Datasets:

Galahad3x

/

QADatasetForPatho

like 0

1400

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

How many people were dead?

1401

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

Which contaminants or viruses or bacteria were found?

1402

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

Which were the symptoms?

1403

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What did the patients have?

1404

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What were the first steps?

1405

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What did they do to control the problem?

1406

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What did the local authorities advise?

1407

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What were the control measures?

1408

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What type of samples were examined?

1409

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What did they test for in the samples?

1410

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What is the concentration of the pathogens?

1411

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What steps were taken to restore the problem?

1412

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What was done to fix the problem?

1413

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What could have been done to prevent the event?

1414

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

How to prevent this?

1415

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What were the investigation steps?

1416

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What did the investigation find?

1417

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

How was the infrastructure affected?

1418

Outbreak of tularaemia in central Norway, January to March 2011

From January to March 2011, 39 cases of tularaemia were diagnosed in three counties in central Norway: 21 cases of oropharyngeal type, 10 cases of glandular/ulceroglandular type, two of respiratory and two of typhoid type. Three cases were asymptomatic and clinical information was unavailable for one case. The mean age was 40.3 years (range 2-89 years). Thirty-four reported use of drinking water from private wells. An increased rodent (lemming) population and snow melting may have led to contamination of the wells with infected rodents or rodent excreta. Outbreak description: From 1 January to 25 March 2011, 39 confirmed cases (16 female and 23 male) of tularaemia were reported from the counties of Sør-Trøndelag (28 cases), Møre og Romsdal (5 cases) and Nord-Trøndelag (6 cases) in central Norway. A confirmed case was defined as a person who had clinical symptoms compatible with tularaemia or had used drinking water from the same source as a previous case, and in whom Francisella tularensis infection was confirmed by a laboratory test as described below. The cases were geographically scattered within each county, involving 13 different municipalities (Figure), and were not linked to one common source. In comparison, seven cases were reported in total from other parts of the country in the same period. In 2009 and 2010 four and eight cases respectively were reported from central Norway. Figure. Geographical distribution of confirmed cases of tularaemia in Norway, January to March 2011 (n=39) Tularaemia is a zoonotic disease caused by the bacterium F. tularensis. Four F. tularensis subspecies are recognised: tularensis, holarctica, mediasiatica and novicida. In Europe, the infection is due to subspecies holarctica which causes in general less severe disease than subspecies tularensis, which is common in North America. Several vectors may be involved in transmitting the disease to humans, commonly rodents and hares, but infection may also be transmitted via insect bites [1]. Several clinical forms are recognised, with oropharyngeal and ulceroglandular disease being the most common clinical presentations in Norway [2]. Oropharyngeal disease is commonly associated with contaminated food and water, while ulceroglandular forms are more often seen when there has been skin contact with infected animals or after insect bites [3]. Outbreaks of oropharyngeal tularaemia have previously been reported from several European countries [3,4]. Tularaemia is a notifiable disease in Norway and during the past 10 years, three outbreaks were reported in Norway [5-7] and all were associated with water sources in areas where dead lemmings (Lemmus lemmus) had been observed previously. From 2001 to 2010, between three and 66 cases of tularaemia were reported annually in the whole country, with an increase from 16 to 32 cases on average (data available from: www.msis.no). This increase may in part be explained by the outbreaks mentioned above. Diagnosis and clinical presentation: In the outbreak described here, the most common clinical presentation was fever and pharyngitis (oropharyngeal type, 21 cases) and cervical lymphadenopathy (glandular/ulceroglandular type, 10 cases). Among the remaining eight tularaemia cases, two were classified as respiratory and two as typhoid type, while three were asymptomatic and clinical information was unavailable for one case. The diagnosis was primarily established by serology (microagglutination and an in-house IgG/IgM Elisa) in 30 patients [8], by F. tularensis specific PCR analysis in seven patients [9] and by blood culture (BactAlert, BioMerieux) in two patients. The two bacterial isolates were verified as F. tularensis by PCR and sequencing of the 16S rDNA gene, and confirmed as non-subspecies tularensis by pdpA PCR [10]. Thirty-four of the 39 diagnosed cases had been drinking water from a private well or a stream. F. tularensis DNA was detected by PCR in filtered water from five different wells tested in Sør-Trøndelag. Seven cases in one municipality were linked to the same water source. Apart from that, only two cases have been confirmed to share a common well so far. Follow-up serology has been recommended for several of the persons exposed to some of the putative water sources. Discussion: The current outbreak involves a large number of municipalities in three counties in central Norway. The clinical presentation with oropharyngeal tularaemia and cervical lymphadenopathy linked to the use of private wells in the winter season makes contaminated water the most likely source of infection in this outbreak. Detection of F. tularensis DNA by PCR analyses in some of the wells supports this assumption for some of the cases. Use of private wells is relatively common in rural areas of Norway although exact data on such use are not available. The precise mechanism of contamination of the wells with F. tularensis is as yet unknown. However, November and December 2010 were unusually cold months, while in January 2011 temperatures increased leading to melting of snow and possible contamination of private wells by surface water contaminated with bacteria from rodent cadavers or rodent excreta. Since the incubation period for tularaemia may be up to three weeks, and time from symptoms until seroconversion might be up to six weeks, more cases may follow. Tularaemia has traditionally been called both ’lemming fever’ and ’hare plague’ and this clearly indicates rodents and hares as transmitters of disease. Years with a great increase in the rodent population are seen with intervals of about three to four years [11] and in the summer and autumn of 2010, a high density of lemmings could be observed in the southern and central parts of Norway. Simultaneously, the Norwegian Veterinary Institute observed a wide geographical distribution of fatal cases of tularaemia in the mountain hare (Lepus timidus) in these regions [12]. The mountain hare is very susceptible to this infection and normally dies from septicaemia within a few days after exposure. The use of small streams and private wells as a source of drinking water and for other purposes in rural areas of Norway is a matter of concern. In existing guidelines issued by the National Institute of Public Health the population is advised to boil drinking water and inspect the wells for dead rodents in case of suspected or confirmed cases of waterborne tularaemia. Every well owner should make the necessary effort to prevent small rodents from entering the well water by carefully covering every opening and plugging every small holes where the rodents can enter. It is also important to secure the well from contamination by surface water after snow melting. In case of proven or suspected contaminated wells, the water should be disinfected before further use. However, this may not be feasible for persons who use drinking water from a stream. The Norwegian Food Safety Authority has recently released information to the media and to the general public with similar advice and information in relation to the current outbreak. The local health authority in each municipality is responsible for instituting infection control measures including advice to the public and investigations of the putative drinking water sources.

What were the infrastructure complaints?

1419

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What happened?

1420

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What was the event?

1421

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

When did this happen?

1422

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

When did this event start?

1423

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What is the date of this event?

1424

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

How long was the event?

1425

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

How long did the event last?

1426

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

In which street did this happen?

1427

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

In which city did this happen?

1428

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

In which region did this happen?

1429

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

In which country did this happen?

1430

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

Where did this happen?

1431

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What caused the event?

1432

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What was the cause of the event?

1433

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What source started the event?

1434

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

How was the event first detected?

1435

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

How many people were ill?

1436

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

How many people were hospitalized?

1437

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

How many people were dead?

1438

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

Which contaminants or viruses or bacteria were found?

1439

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

Which were the symptoms?

1440

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What did the patients have?

1441

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What were the first steps?

1442

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What did they do to control the problem?

1443

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What did the local authorities advise?

1444

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What were the control measures?

1445

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What type of samples were examined?

1446

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What did they test for in the samples?

1447

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What is the concentration of the pathogens?

1448

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What steps were taken to restore the problem?

1449

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What was done to fix the problem?

1450

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What could have been done to prevent the event?

1451

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

How to prevent this?

1452

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What were the investigation steps?

1453

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What did the investigation find?

1454

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

How was the infrastructure affected?

1455

Searching for a water supply connection in the Cabool, Missouri disease outbreak of Escherichia coli 0157H7

Abstract--A recent disease outbreak resulting in 4 deaths, 32 hospitalizations and a total of 243 documented cases of diarrhea was linked epidemiologically and by on-site data gathering supported by the use of a distribution system model to the public water supply. The pathogenic agent, Escherichia coil serotype 0157:H7, was isolated from patients' fecesin tests conducted by the Centers for Disease Control. Illness was restricted to people using public water supply. Untreated groundwater quality was not a factor but some disturbances in the distribution system, possibly 43 water meter replacements and 2 line breaks, may have allowed contaminants to enter the water supply. This is the first time a distribution system model has been used ~o show that the pattern of illness occurrences in a waterborne outbreak study could be related to water movement patterns in the distribution network. Key words---outbreak, Escherichia coli 0157:H7, distribution system model, public supply, untreated groundwater INTRODUCTION Cabool, Mo. (population 2090) is located in the Ozark hills, near the border with Arkansas. The area is in a limestone formation and sink holes are part of the topography. Dairy agriculture forms the economic base for this farm community. Before the outbreak the public water supply for Cabool was obtained from deep wells and then transmitted untreated to storage reservoirs in the distribution system. During the period 15 December 1989-20 January 1990, residents and visitors to the community of Cabool, Mo., experienced 243 cases of diarrhea (85 bloody) and four deaths (Swerdlow et al., 1992). The Centers for Disease Control (CDC) conducted a household survey from which they concluded that persons living inside the city (on the municipal water) were 18.2 times more likely to develop bloody diarrhea than for persons living outside the city using private well water. After a boil water order was issued (5 January 1990) in the city of Cabool, the number of new cases rapidly declined. The city authorities implemented a chlorination program for the community water supply on 12 January. The majority of the cases occurred during a period of exceptionally cold weather during which there were numerous water meter replacements (14-22 Dec. 1989) on service lines and two breaks in the water distribution lines (22-23 Dec. and 25-26 Dec. 1989). Escherichia coil serotype 0157:H7 was found in the feces of some infected individuals. This organism has been detected in most areas of the U.S.A. (Ostroff et al., 1991) and is reported to be a common cause of bacterial diarrhea in Canada and Great Britain (Griffin et al., 1988; Laboratory Center for Disease Control, 1987; Public Health Laboratory Service, 1987). Two recent studies from the U.S.A. have reported that this organism is a more common cause of diarrhea than Shigeila (MacDonald et al., 1988; Marshall et al., 1990). Very young and very old persons are most likely to become ill and most likely to develop complications (Stewart et al., 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et al., 1987). One serious complication of E. coli 0157:H7 infection is the hemolytic uremic syndrome. The infective dose for this pathogenic E. coli is estimated to range from 10 to 100 organisms with infectivity being most severe among infants, senior citizens and patients in nursing homes (Stewart et aL, 1983; Ryan et al., 1983; Pudden et al., 1985; Krishnan et ai., 1987). In the Cabool outbreak, elderly persons were more likely to become ill and the four deaths were among elderly citizens. Previous outbreaks of E. coli 0157 have been caused by contaminated hamburgers (Griffin et ai., 1988). To our knowledge this is the first outbreak associated with waterborne transmission. Another outbreak associated with waterborne transmission of this pathogen occurred in Scotia,! duriitg the Summer of 1990 and has been reported (Dee et al., 1991). This E. coli serotype has been isolated from the feces of healthy dairy cattle, suggesting that cattle are an important reservoir for this pathogenic agent (Martin et al., 1986; Borczyk et al., 1987; Orskov et al., 1987; Wells et al., 1992). Cattle raising is a major agricultural activity on the Cabool watershed. In general, the prevalence of various E. coil serotypes pathogenic to man in farm animals has been reported to be less than 9% (Oeldreich, 1972). Investigative approach At the invitation of the CDC (U.S. Public Health Service), the Missouri State Health Department and the Missouri Department of Natural Resources, staff members of the U.S. Environmental Protection Agency (EPA), Drinking Water Research Division (Cincinnati, Ohio) went to Cabool to review data, conduct a sanitary survey and collect additional water samples to bc sent to the EPA Research Center in Cincinnati for further study. Since this investigation of the water system was conducted 4 weeks after the main impact of the outbreak had subsided, the strategy for investigating the possible involvement of water supply focused on the study of long term monitoring data from the groundwater supply (aquifer and distribution quality). In addition, attempts were made to model the distribution system for movement of water through the system. A general inspection of the drinking water supply system and operating practice was also conducted. Capturing a segment of the water representing water quality remaining in the distribution system at the start of this investigation but from the period of the outbreak did not appear probable because of time elapsed (6 weeks). However, there was a remote chance that some water from the outbreak period might still be found at the extremities of the distribution system where water usage was low. METHODS Specialwater samples of 200 ml volumewere collected on 14 and 15 February 1990 from the distribution network in addition to sample collections at the two operating wells (wellNos 5 and 6) for analysesat the EPA Research Center, Cincinnati. All of the special sampling sites were selected from the extremities of the distribution system where water flow might be static and perhaps representative of water from 4 to 6 weeks prior. These water samples were analyzed for total coliform bacteria by the membrane filter method using m-Endo LES agar and m-T7agar. Sampleswerealso analyzed by multiple tube procedure using lauryl tryptose broth with confirmation in brilliant green bilebroth and by the Colilert system (EnvironetiesInc., Branford, Conn.). Total coliform isolates were identified to species using the API 20E multiple test system (Analytab Products, Plainview, N.Y.). Heterotrophic plate count analyses were conducted by the pour plate procedure using plate count agar and by the spread plate procedure using R2A agar. All microbiologicalanalyses were conducted according to standard procedures (APHA, 1989). Survival experiments were conducted using filter-sterilized (0.22#m porosity Duropore® filters, Millipore Corp., Bedford, Mass.) water. An 18-h culture of E. coil 0157:H7 grown at 37°C was washed twice by centrifugation using Standard Method~ phosphate buffer and used as the inoculum. The samples were held in the dark at 5°C and assayed L~eriodically by the spread plate procedure using sorbitol MacConkey agar. Chlorine determinations Chlorine levels were measured in the field using the N,N-diethyl-p-phenylenecfiatrdne (DPD) colorimetric method (APHA, 1989). Samples Samples were collected in sterile containers to which sodium thiosulfate was added to neutralize up to 5 mg/l of any disinfectant residual (APHA, 1989). Samples were shipped on ice by overnight carrier and analyzed within 24 h of collection. Source water quality The drinking water system in Cabool, Mo. is supplied by four municipal wells. Figure 1 shows the distribution network, the two major wells serving the system and the water supply storage tanks. Prior to the outbreak, no disinfectant was added to the municipal water supply. A local dairy industry uses water from both the municipal system and from its own private wells (D in Fig. 1). Although there are four wells within the public water supply system, only well Nos 5 and 6 were operating during the time in question. Two other wells (Nos 3 and 4) are used only during the summer months when the demand for water is high. During the winter months, well Nos 3 and 4 are valved off from the system and electrical power to the pumps is disconnected. Well No. 6 (which supplies approx. 55% of the water supply) is located near the southwest perimeter of the distribution system. This well generally operates continuously. The well depth is approx. 305 m (1000 it) and is cased to a depth of 135 m (450 it) where a submersible pump 7.21/s (115 gpm capacity) is located. Well No. 5 is located along the North perimeter and operates intermittently as demand requires. Typically, during winter, well No. 5 operates for approx. 8 h/day though this use period varies considerably With demand. It is a drilled well approx. 396 m (1300 ft) deep and cased to a depth of 134.4 m (441 it). The submersible pump in well No. 5 is set 134m (440 it) below ground and has a capacity of 17.3 l/s (275 gpm). Both well heads are housed in small buildings constructed atop poured concrete slabs. Neither well head appeared to be susceptible to surface run-off from agricultural fields on the watershed. The well head structures housed the sanitary seal around the wells, totalizer flow meters and fluoride injection systems to maintain desired fluoride levels in the distribution water. Drilling logs were not available for well No. 6 but the logs for well Nos 3, 4 and 5 indicated that the underlying geological formations were very similar to each other, possibly being in the same aquifer. Overlying geologic formations included red clay and gravel near the surface (to a depth of approx. 150 it) followed by limestone, sandstone and flint formations. The wells were cased continuously from the surface to a depth of at least 122m (400 ft) and passed through various limestone formations described as very hard, medium hard and hard in the drilling log. The drill logs indicated that the aquitard layers were fairly impervious to surface water infiltration. Monitoring data supplied by the Missouri Department of Natural Resources for the period 9 November 1981-1 l January 1990 for city well Nos 5 and 6 revealed that no coliform bacteria were ever detected using the membrane filter test on 100 ml sample portions. As a further check on water quality in the aquifer, permission was given by the Mid American Dairy Company (D in Fig. l), to examine their records on the three private wells (located on plant property) they used in milk processing. These wells were drilled to the same depth as the city wells (approx. 305 m) and therefore assumed to be in the same aquifer. Once each month one of the three wells is examined by the certified milk laboratory for coliform bacteria, using the same membrane filter technique employed by the certified State Health Department Laboratory. Inspection of laboratory reports for the years 1987-1990 reveal no coliforms present in 100 ml test portions analyzed for any of the three private well water supplies. These findings suggest that the aquifer was protected from surface water contamination. Water supply storage Water storage was provided by two storage tanks: a 1900m 3 (500,000 gal.) tank (T500 in Fig. 1) and a 230m 3 (60,000 gal.) elevated tank (T60 in Fig. 1). A third elevated water storage tank, 380 m3(100,000 gal.), designated as YT in Fig. I, was not being utilized. Based on discussions with local personnel there is little draw-down (less than 0.6 m/ day) in the water storage tanks. This observation was later verified by hydraulic analysis. Water storage tank T500 is located approx. 180 m (600 r) east of well No. 5 in the northwest part of the distribution system. This storage tank is fed primarily by well No. 5. According to utility personnel the pump at well No. 5 is turned on when the water level in tank T500 drops to 441 m 0447 it) or below and continues pumping until the water level reaches approx. 441.6 m (1449 ft) above sea level. The upper water level is about 2.4 m (8 ft) below the overflow level. The elevated storage tank (T60) is centrally located in the distribution system and the overflow elevation for this storage tank is approx. 444.1 m (1457 it) above sea level. Water from both wells feeds this storage tank at different times during normal operation. The yellow storage tank (YT) is an elevated tank and is located less than 1.6 km (1 mile) west of well No. 6 and is at the southwest end of the distribution system. A faulty pressure valve has prevented this storage tank from operating properly. The ground elevation of 396 m (1300 ft) is such that the water level will only reach the base of the elevated storage compartment when the overflow elevations of the other two storage tanks are reached. Thus, this storage tank provided storage only in the column rising up to the compartment (bulb) and could be a substantial deadend. All of the storage tanks in the system use a single riser tube for both filling and drawing. Therefore, when the tanks were being used, the last water placed in the tanks was the first to be used. Thus, the water in storage could be stratified and provide an excellent location for bacterial activity. Distribution system integrity On-site investigation of the distribution system reveals the pipe network consists of a mixture of cast iron, galvanized iron and plastic pipe with sizes of 5.08, 10.16, 15.24, 20.32 and 50.8cm (2, 4, 6, 8 and 10in.) diameter intermixed. Meter records also indicated a 35% level of unaccounted for water. This could have occurred through leaking mains, inaccurate meters or non-metered use. The dairy has a backflow prevention device that isolates the private well supplies from the Cabool distribution system. There are no records on water pressure in the system but a pressure gage on the water supply side of an alarm valve to the fire control system at Mid Am Dairy reveals pressure on 14 February 1990 was fluctuating around 7.75 kg/cm2 (110 psi) (Fig. 1). The water pressure on 22-23 December 1989, during a major loss of water over several hours from a main break, could not be determined beyond a note of reported "low water pressure" in the neighborhood. A second line break occurred on 26 December resulting in similar duration and water loss. Figure 1 shows the location of both line breaks in the central part of the system and the meter replacements. Discussion with the Cabool water plant superintendent indicated there is a general practice to flush all line repairs by turning on the valve at the lower elevation of the break area and flushing the line out a fire hydrant at the higher elevation for 15min. Line disinfection is not practiced and there has never been an annual scheduled flushing of the entire water distribution system since water quality was considered good and there were no complaints of taste and odors. Based on customer recollections, the two major line breaks on the system did not reduce water pressure systemwide although localized low water pressure created opportunities for back siphonage near the break and a pathway for sewage infiltration. Line flushing around the breaks was a 15-minwater release at nearest fire plug below the repair and was not preceded by pipe disinfection. No follow-up bacteriological sampling was conducted, and the routine monitoring schedule did not allow for sampling during this time frame. Surface water run-offfrom the watershed did infiltrate the separate sewage system in addition to drainage through an open culvert system along the city streets. As a consequence, sewage and stormwater run-off may be found near water pipes, meter boxes and service lines, particularly after a major storm event. Sincethe weather prior to and during the outbreak was severely cold, stormwater run-off was probably not a factor except where surface water collected in servicemeter boxes. Sewage, however, being at and above freezing temperature would continue to saturate adjacent soil as it flowed through the deteriorated pipe collection system towards the overloaded wastewater lagoon in the southeastern part of town. Overflow from the lagoon runs over low-lying land that also contains distribution pipes before reaching the river nearby. The sewage collection system in Cabool is located (for the most part) away from the drinking water distribution lines, but does cross or approach the water lines in several locations. Even though there had been no rain for several days during the site visit in February, severalmanhole covers showed indications of overflow problems. Various sewage paper products littered the areas around a few manhole covers, and several other entry structures showed small erosion gullies around their periphery. The final manhole or collection box before entering the wastewater treatment lagoons showed evidence of routine overflow. In fact, there were periodic overflows during the time it took to walk around the lagoons. This overflow ran over land to the Big Piney River. A water main ran directly underneath this overload waste flow. RESULTS State agency monitoring data on the distribution system for the period of 5 December 1989-12 January 1990 revealed seven coliform positive samples among 65 distribution samples collected. None of these official routine or special samples were collected between 19 December and 2 January. However, a dairy inspector, who lives across the street from the first main break collected a water sample in his home on 8 January (flaming the faucet and using a sterile sample bottle) and transported it to the certified dairy laboratory in St Louis for analysis on the same day. The laboratory reported 22 coliforms per 100 ml in the water sample. Two follow-up samples taken 11 January 1990 at the home of the dairy inspector before chlorination of the water system was instituted, were negative for coliforms. Results of the bacteriological analyses (Table 1) reveal no coliforms were detected from well No. 5, well No. 6 and the two distribution sites (Grandview Terrace and Rt 60 at M Highway). However, 55-95 coliforms were detected in three distribution sites from the southwestern part of the pipe network and in water adjacent (100 m distant) to the yellow storage tower, indicating that there had been a contaminating event before the February sampling period. On l0 July 1990 the Cabool distribution system was again sampled at the yellow water tower and two areas of potential slow flow (Kalco Manufacturing near the yellow tower and Cedar Bluff sites at the extreme southeastern portion of the system). While there were less than 1.1 coliforms per 100 ml detected at the latter two sites, the yellow water tower sample did yield a slow fermenting coliform (l.l organisms/ 100 ml) identified as Klebsiella oxytoca by the API- 20E species differentiation system. The heterotrophic plate count at this site was 1,000,000 organisms per ml, suggesting a static water situation. Since the water supply is now disinfected with chlorine, residual surviving coliforms in the slow moving water around the yellow tower area may have still persisted but in a stressed state. On 12 January 1990, the city began chlorinating the well water by discontinuing fluoridation, and adding a liquid commercial bleach to the water supply using the fluoride pumps. Free available chlorine was monitored at a sampling port located 30 ft from the injection point at well No. 5 and another one located 150 ft from the well No. 6 injection point. The measured free chlorine concentration at the two sampling ports for the first 31 days after chlorination began is shown in Fig. 2. The chlorine dose applied at well No. 6 was less variable than well No. 5 due to the fact that well No. 5 is only operated as needed and well No. 6 is pumping constantly• There still was a highly variable measured amount at both locations. Coliform colonies detected from the three positive samples were submitted to purification on plate count agar and then identified to species using a commercial multi-test system (API-20E). These results are shown in Table 2. While the profile of coliform species may have been biased since only 15 or 16 colonies from each sample were examined, all three coliform positive samples contained Escherichia hermanii, a possible fecal organism (Brenner et al., 1982). Although E. hermanii is not known to cause gastroenteritis, its presence is significant because this organism closely resembles E. coli 0157:H7 in its biochemical profile and has been found in raw milk, ground beef and feces (Lior and Borczyk, 1987)• Further study using an enrichment process revealed that some of these coliform isolates were tetracycline resistant, a characteristic shared with the outbreak strain of E. coil 0157:H7. An additional coliform (Klebsiella pneumoniae) was isolated in a secondary study of these samples and grew at 44.5°C, fitting the definition of a fecal coliform. All Enterobacter sp. isolates were found to be resistant to cephalothin, tetracycline and ampicillin. The E. hermanii isolates were resistant to carbenicillin and showed intermediate resistance to ampicillin. The identical antibiotic resistance patterns seen for the same species of coliform bacteria isolated from the three different locations in the distribution system suggests that the organisms originated from a common source of contamination. Since infections with E. coil 0157:H7 occurred over a 2-3 week period, persistence of E. coli serotype 0157:H7 was an important aspect of this investigation. To study this aspect, water from well Nos 5 and 6 and two distribution sites of slow flow were filter sterilized and inoculated with a strain of E. coli 0157:H7 (strain A) obtained from the University of Wisconsin, Food Microbiology Department. Later when the specific pathogenic strain isolated from patient feces became available, a new sample from well No. 5 was again filter sterilized and inoculated with the specific pathogen strain (strain B). A high density inoculation ranging from 540,000 to 1,800,000 cells per ml was used in anticipation of a fast decline in E. coli density. An incubation temperature of 5°C was selected to simulate what may have been the water temperature in the distribution system at the time in question. Results of these survival experiments are given in Table 3. This information indicates that both strains of the E. coli serotype had a relatively slow rate of die-off so that after a week or more, relatively high concentrations could have remained. In fact, there was only a 2 log decline in the pathogenic strain in 5°C well water after 35 days. By contrast, parallel cultures of either strain of this pathogen held in Cabool water at 20°C revealed a 5 log decline after 35 days of storage. SYSTEMS M O D E L I N G In an attempt to gain insight on how system failure and/or contaminant propagation can influence water quality, a model developed by EPA's Drinking Water Research Division was applied to the Cabool, Mo. water supply system. The model and the approach utilized in this study had been tested and validated extensively but never before applied in a waterborne disease investigation (Clark et al., 1988; Males et al., 1988; Grayman et al., 1988, Clark and Coyle, 1990). Both steady-state and dynamic modeling approaches were attempted in this investigation. The steady-state modeling (assuming uniform demand and supply throughout the system) was used to determine where the water from well Nos 5 and 6 would be expected to be found under "normal" or average cold weather demand conditions. Dynamic modeling (variations in system conditions over a given period) was used to track contaminant propagation in the system. These conditions are described later. In this case study, the distribution system was represented by a link-node network. A node is created along a pipe when there is a major change in pipe direction, an intersection or tee, or a change in pipe diameter. Wells, tanks and major water users are also considered nodes. The run ofpipe between two nodes is then considered a link. A hydraulic model was then used to determine flow directions and velocitiesin links. Figure 3 shows the information contained in all the previous figures with additional data on households where cases occurred. Based on the previous analysis it seems unlikely that the public water sources (well Nos 5 or 6) or a possible dairy interconnection caused the outbreak. It seems more likely that the outbreak resulted from disturbances in the system that are in close proximity to most of the outbreak cases. The hypothesis that some disturbance in the system allowed contaminants to enter and be propagated throughout the distribution system was therefore pursued. Meter replacement scenario Movement of water and contaminants resulting from hypothetical contamination at sites where water meters were replaced in December 1989 was studied. Hydraulic patterns associated with the normal conditions and break conditions were used in the analysis. For each situation, the water at each node in the vicinity of the meter replacements was contaminated by assuming an initial arbitrary concentration of l0s organisms per ml at the site with no die-off. No further additional contamination was added. The movement of the contaminated water was then traced and nodes that would receive the contaminated water (at various dilution levels) were identified and plotted. The contaminant propagation varied widely. Figure 3 shows that in some meter replacements the spread of the contaminant covered a large section of the distribution system, but in other cases it remained localized. Since only three homes that had meters replaced had illnesses, it was concluded that meter replacement was not the major cause of the outbreak but could have accounted for the early cases prior to the line breaks. Main break and repair scenario A dynamic analysis of the movement of water under normal and break conditions was simulated. EPA's Dynamic Water Quality Model (DWQM) was applied to examine the movement of flow in the system under the normal operating conditions prior to the break being repaired and hydraulic situations simulating recovery following repair. A conservative contamination level of 105 organisms per ml in a 0.6 l/s (10 gpm) flow for a period of 4 h of continuous flow to match the normal hydraulic demand in the area was assumed at each of the breaks. No die-off of organisms was assumed. Movement of water and contaminants resulting at each of the two break sites (22-23 Dec. and 25-26 Dec.) was simulated using the hydraulic conditions immediately following the repair of the breaks (Figs 4 and 5). Figures 4 and 5 show the extent of the flow of contaminated water resulting from the two breaks. Note, flow of contaminated water overlays most of the outbreak cases with at least 10-100 organisms (4 log reduction) still present. Combining both break patterns provided an overlay of 85% of all household case locations. Because several hours had elapsed before the breaks were repaired, the tanks had been drawn down quite extensively. Thus, it required nearly 36 h of continuous operation of both wells for the tanks to recover. This scenario resulted in well No. 5 operating to fill the large 1900 m3 (500,000 gal.) tank while well No. 6 served the daily demand and reaching portions of the system not normally receiving well No. 6 water. This would enable contaminated water from both break areas to cover an extensive area, exposing nearly the entire service area to contaminated drinking water. DISCUSSION The field investigation revealed the need for several changes in operational maintenance of the distribution system, monitoring site selection on the pipe network and infrastructure improvements in sewage collection and its treatment. The concern was that any disturbances in the distribution network may provide a contamination pathway from stormwater run-off or sewage infiltration. The distribution system must protect the quality of water transmitted throughout the system. In the case of the Cabool, Mo. water system, water supplied to consumers was not disinfected prior to the outbreak, so residual disinfection was not available to provide a measure of protection from contaminants that might enter through line breaks, back siphonage or crossconnections. Flushing the entire distribution system in a systematic manner to get more movement of the chlorine residual into all parts of the pipe network, was needed to remove static water from slow flow sections, deadends and stratified water in storage tanks on a periodic basis. Another area of concern was the water service meter repair practices that do not include disinfection application. A rigorous protective protocol must be followed during the repair or replacement of existing mains and service meters in order to avoid bacteriological contamination of the distribution network (Buelow et al., 1976). No disinfectant can possibly be effective when lines contain sediments that provide a protective habitat for bacterial growth. Pipe interiors, meter fittings and valves must be protected against contamination. Meter boxes should be drained of surface water seepage prior to meter replacement and new meters carefully inserted free of soil particles. After completion, lines should be flushed at a minimum velocity of 76.2 cm/s (2.5 ft/s). In the real world of a mixed population of organisms, antagonistic competition would have depressed the persistence rate of E. coil serotype somewhat but the laboratory experiments with Cabool groundwater still suggest that £. coil serotype 0157:H7 was capable of persisting long enough to reach a significant portion of the distribution system. While the pathogenic agent was never detected in the groundwater, there is reason to believe the coliform isolates from special sampling and the causative agent were at one time closely associated, being indicative of contamination from the sewage system or storm drains. Regardless, circumstantial evidence strongly suggests that a break in the public health barrier concept did occur between sewage, stormwater and water supply. For example, six cases of bloody diarrhea were identified as having occurred prior to the first water main break but after 43 meter replacements on the system. Seven other cases were reported between the two water main breaks that were 3 clays apart, with the remaining 72 cases identified within a week of the second break. This situation points to the possibility that E. coli 0157:H7 was prevalent for several weeks in the community. These observations suggest that the existing sanitary sewer system was prone to infiltration from stormwater run-off and underdesigned for the capacity transmitted. Inspections were not made as to possible sewer line collapses or other blockages that would amplify the collection and treatment problems. What was considered as a consequence was the fact that sewage overflows transgress surface areas over drinking water distribution lines and in a few locations run across sites where water meter boxes are located. There are some unique characteristics for this pathogen that should be recognized by state public health and water supply authorities monitoring municipal water supplies. The organism lacks the enzyme /~-giucuroniclase, will not grow at 44.5°C and may give variable gas fermentation results with lactose at 35°C (Doyle and Schoeni, 1984; Kirshnan et al., 1987; Hartman, 1988). As a consequence, the organism will not produce a fluorescence in the 4- methylumbelliferyl-/J-D-glucuronidase (MUG) assay (Chang et al., 1989) and will not grow in any fecal coliform test using elevated temperature incubation. The epidemic strain was routinely MUG negative using lauryl tryptose broth and EC broth containing MUG and also in the Colilert and Coliquik commercial media. The organism does differentiate as sheen colonies on a conventional total coliform membrane filter M-Endo type medium but will not produce a positive result in the multiple tube or P-A fermentation tests for coliforms. Thus, the occurrence of E. coil 0157:H7 may be somewhat difficult to identify in routine monitoring for E. coil as fecal coliforms. If sewage or surface water drainage was the origin for this pathogenic E. coil, then the question arises as to why this organism and other coliforms were not detected in the contaminated water supply. It is most important to note that no official monitoring of the public water supply was done during the outbreak period. One special sample was collected at a home across the street from the first line break on 8 January 1990. The certified laboratory reported the sample contained 22 coliforms per 100 ml but no analysis was done on the sample for fecal coliform or £. coll. The next sampling at this same site was done on 11 January 1990 and tested negative for coliform bacteria. By the time the contamination had passed through the system disinfection of the water supply had begun. The other weakness in the monitoring program was to focus site selections to a few locations near the center of town. Perhaps if the sampling locations had included areas near deadends, some indication of the contaminating event would have been detected during the month. Distribution sample site selection should not focus only on locations in the center of town. More effort should be made to vary locations over the year so that water samples are frequently collected from the periphery of the pipe network. At these areas of slow flow there are more opportunities to capture water quality changes that could alert the water operator to the need for line flushing so that a chlorine residual could be restored or bacterial growth in sediments suppressed. The proper and continual addition of a disinfectant needs to be practiced at all well heads. The data collected by the system operators have shown that a consistent chlorine residual is not maintained at the first sampling tap after chlorine addition. To help regulate the chlorine dose applied to the water system, the system operator should install effective chlorination equipment. This would mean installing either a liquid or a gaseous chlorine feed system. Liquid sodium hypochlorite (12-15% available chlorine; NOT HOUSEHOLD BLEACH) can be metered into the pipe systems at the well heads and with proper injectors/diffusers, no additional mixing would be required. Chlorine could also be injected at the well heads using a gaseous chlorine injection system. Once the proper equipment is installed, the system should be monitored to assure that a constant dose of chlorine is added to the water system. All deadends should be flushed until a chlorine residual is obtained in the flush water. Routine monitoring for chlorine residual near the first customer locations for each well, various sites within the distribution system and at taps at the extremities of the distribution system need to be done in order to assure that a proper chlorine residual is available throughout the distribution system. The maintenance of a chlorine residual in the distribution system would help to insure the integrity of the water supply in case of future contamination within the system. Infiltration of stormwater run-off into the sewage collection system caused frequent overflows of wastewater from manhole covers over junction boxes. The pipe network needs infrastructure revitalization to stop surface surges of raw sewage. The wastewater lagoon system overflows periodically as a result of stormwater infiltration spilling over the banks in the vicinity of a section of the distribution line, before reaching a small stream. One of the retention lagoon cells was reported to be off-line which may reduce retention time and treatment process effectiveness, indicating operational practices need revision. CONCLUSIONS Those professionals familiar with the investigation of disease outbreaks know that establishing cause and effect is a difficult task. The principal value in reporting this type of research investigation lies in the following points: (1) An outbreak of gastrointestinal illness attributed to E. coli 0157:H7 occurred in a groundwater supply that had been historically characterized as being of excellent quality and not in need of disinfection. (2) This was the second reported waterborne occurrence of E. coil 0157:H7. This time in a community with an outbreak of 243 cases of bloody diarrhea and 4 deaths. (3) Laboratory characteristics of the organisms are not typical of the classical E. coll. The organism is not detectable at 44.5°C in either the multiple tube or membrane filter fecal coliform tests, nor will it produce fluorescence in various MUG based media at 35°C. (4) Illness cases were restricted to people using public water supply. Intensive search by CDC into other likely sources of contamination (meat, milk and sewage aerosols) proved negative for this pathogenic agent. (5) This investigation provided the first opportunity to use a distribution system model to study the pattern of illness occurrences in relation to normal water movement patterns in the distribution system, develop descriptions in water flow caused by line breaks and map the diffusion of a pathogenic agent through either line breaks or meter replacements.

What were the infrastructure complaints?

1456

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What happened?

1457

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What was the event?

1458

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

When did this happen?

1459

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

When did this event start?

1460

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What is the date of this event?

1461

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

How long was the event?

1462

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

How long did the event last?

1463

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

In which street did this happen?

1464

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

In which city did this happen?

1465

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

In which region did this happen?

1466

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

In which country did this happen?

1467

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

Where did this happen?

1468

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What caused the event?

1469

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What was the cause of the event?

1470

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What source started the event?

1471

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

How was the event first detected?

1472

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

How many people were ill?

1473

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

How many people were hospitalized?

1474

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

How many people were dead?

1475

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

Which contaminants or viruses or bacteria were found?

1476

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

Which were the symptoms?

1477

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What did the patients have?

1478

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What were the first steps?

1479

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What did they do to control the problem?

1480

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What did the local authorities advise?

1481

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What were the control measures?

1482

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What type of samples were examined?

1483

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What did they test for in the samples?

1484

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What is the concentration of the pathogens?

1485

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What steps were taken to restore the problem?

1486

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What was done to fix the problem?

1487

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What could have been done to prevent the event?

1488

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

How to prevent this?

1489

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What were the investigation steps?

1490

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What did the investigation find?

1491

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

How was the infrastructure affected?

1492

Two Drinking Water Outbreaks Caused by Wastewater Intrusion Including Sapovirus in Finland

Abstract: Drinking water outbreaks occur worldwide and may be caused by several factors, including raw water contamination, treatment deficiencies, and distribution network failure. This study describes two drinking water outbreaks in Finland in 2016 (outbreak I) and 2018 (outbreak II). Both outbreaks caused approximately 450 illness cases and were due to drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. In both outbreaks, the sapovirus was found in patient samples as the main causative agent. In addition, adenoviruses and Dientamoeba fragilis (outbreak I), and noroviruses, astroviruses, enterotoxigenic and enterohemorragic Escherichia coli (ETEC and EHEC, respectively) and Plesiomonas shigelloides (outbreak II) were detected in patient samples. Water samples were analyzed for the selected pathogens largely based on the results of patient samples. In addition, traditional fecal indicator bacteria and host-specific microbial source tracking (MST) markers (GenBac3 and HF183) were analyzed from water. In drinking water, sapovirus and enteropathogenic E. coli (EPEC) were found in outbreak II. The MST markers proved useful in the detection of contamination and to ensure the success of contaminant removal from the water distribution system. As mitigation actions, boil water advisory, alternative drinking water sources and chlorination were organized to restrict the outbreaks and to clean the contaminated distribution network. This study highlights the emerging role of sapoviruses as a waterborne pathogen and warrants the need for testing of multiple viruses during outbreak investigation. Keywords: waterborne outbreak; enteric viruses; contamination; drinking water; wastewater; sapovirus; microbial source tracking; fecal indicators; Dientamoeba fragilis 1. Introduction The drinking water contaminated with pathogenic microbes may cause large community outbreaks with up to thousands of illness cases in both developing and developed countries. Several factors may cause a drinking water outbreak. Raw water contamination, treatment deficiencies, and distribution network failure are among the most common causes [1]. In addition, waterborne outbreaks have been associated with climatic conditions, especially with increased precipitation and heavy rainfall events [1–4]. The source of the contamination is most commonly wastewater which may harbor a large number of diverse pathogenic microbes. In Finland, a food and waterborne outbreak surveillance system has revealed several waterborne outbreaks every year since 1997. In these outbreaks, norovirus has been the most common causative agent followed by Campylobacter [5,6]. In addition to noroviruses, the potential waterborne spread of other enteric viruses, such as adenoviruses [7,8], sapoviruses [9,10], enteroviruses [8], astroviruses [11] and rotaviruses [8] have been reported in Finland. Sapoviruses are close relatives to noroviruses and the clinical symptoms of sapovirus gastroenteritis are indistinguishable from those caused by noroviruses. Though, in general, the clinical severity of sapovirus-associated disease is milder than that for norovirus and rotavirus [12]. Sapoviruses are common in wastewater [13,14], and due to the availability of improved methodologies, these viruses are also now being analyzed and detected more often. An increasing number of reports related to outbreaks and sporadic cases caused by sapovirus have been described, highlighting the emerging role of sapoviruses as a public health concern [15–21]. Traditionally, the microbiological quality of drinking water has been estimated by using fecal indicator bacteria (FIB), such as Escherichia coli, intestinal enterococci and Clostridium perfringens. These FIB are part of the normal flora in the intestinal tract of humans and other warm-blooded animals, and thus they are consistently present in wastewater. However, the capability of these indicators to measure water quality and predict waterborne outbreaks has been questioned [22–24]. Therefore, more specific and sensitive fecal indicators of water quality have been explored. Potential candidates are the genetic markers from the group of Bacteroidales, such as general Bacteroidales genetic marker (GenBac3) [25] and the host-specific HF183 marker [26], used as targets in quantitative PCR (qPCR) assays for the detection of fecal contamination and human wastewater pollution, respectively. Although the qPCR assays are often designed to target the ribosomal RNA gene (rDNA), it has been proven that the detection frequency of fecal bacteria in water can be enhanced by targeting the assays to rRNA transcripts instead of rDNA [27,28]. While Bacteroidales assays are widely applied in studies of microbial source tracking (MST) in surface waters [29], their use as part of community-wide waterborne outbreak investigations is rare [10]. Thus, more data to assess the suitability of these new indicators as a tool to describe drinking water contamination episodes, to detect drinking water quality deficiencies and their application in processes securing good drinking water quality, is needed. This study describes two waterborne outbreaks both caused by the intrusion of wastewater into a drinking water distribution system due to pipe breakage. Causative agents of outbreaks were determined through investigations of patient and water samples and the suitability of both traditional FIB and new candidates (GenBac3 and HF183) to provide water quality information was evaluated. 2. Materials and Methods 2.1. Outbreak Descriptions and Samples This study describes two drinking water outbreaks in Finland in October 2016 (outbreak I) and January 2018 (outbreak II). Both outbreaks were initially caused by the drinking water pipe breakage and subsequent wastewater intrusion into the distribution system. Information regarding the outbreaks was collected from the local investigation reports, including retrospective questionnaires, and personal communications. The outbreaks were defined as waterborne outbreaks with a strong strength of association based on classification criteria presented previously [30,31]. 2.1.1. Outbreak I In outbreak I, the cause of the contamination was a maintenance well containing the air release valves of both drinking water and wastewater pipes (Figure 1). The air release valve of the wastewater pipe allowed wastewater to leak and accumulate into the maintenance well. Due to pipe breakage on the road construction site on 12th October 2016, the under pressure in the drinking water network caused the wastewater inflow from the maintenance well through the air release valve into the drinking water distribution system. The pipe breakage was detected and repaired immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. immediately but the cross-contamination in maintenance well was detected only after six days on 18th October 2016. Drinking water originating from the groundwater source was flocculated with KMnO4, pH was adjusted with NaOH followed by clarification and sand filtration through three sand basins and finally UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for two months (from 16th October to 16th December 2016) and chlorination for 50 days (from 16th October to 5th December 2016). The target chlorine levels were as follows: first 2 mg/L for 3 days, then 4.5 mg/L for 3 days and finally 1 mg/L for 44 days. In addition, alternative water sources were arranged for the water users during the outbreak. The drinking water contamination affected approximately 790 people. In order to estimate the magnitude of illness, questionnaires were sent to the households of the contaminated area. The response rate was 62% (294/471 households). In the analysis, only one response per household was included. Thus, in total, 115 symptomatic cases of 283 respondents were observed (Figure 2a). When respondents’ family members with gastrointestinal illness were taken into account, the estimated number of patients was 458. According to a questionnaire study, the first patients appeared one day after the pipe breakage. The median duration of the symptoms was one to two days and the most frequently reported symptoms included abdominal pain (94%, 101/107), nausea (91%, 100/110), diarrhea (89%, 100/112), abdominal swelling (83%, 86/104), muscular pain (66%, 64/97), vomiting (53%, 52/98) and fever (46%, 42/91). The symptoms suggested a viral point source outbreak with a rapid increase of cases followed by a fast decrease after the mitigation actions (Figure 2a). In the acute phase of the outbreak, stool samples were collected from patients between 19th October and 3rd November 2016, and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed in local environmental laboratories from the water samples collected between 15th October 2016 and 27th January 2017. Drinking water samples were taken for pathogen analyses only after start of the chlorination on 24th October 2016 (n = 1) and 26th October 2016 (n = 3). Pathogen analyses for the water samples were selected and prioritized based on results from patients taking into account the available water volume. The early water samples were analyzed only for sapoviruses and protozoans (Cryptosporidium spp., Giardia lamblia, Entamoeba histolytica, and Dientamoeba fragilis). On 28th November 2016, a raw water sample (dead-end ultrafiltration, DEUF), drinking water samples (n = 3, DEUF), biofilm samples from water meters (n = 9) and a sample from the contamination site (maintenance well) were collected and analyzed for sapo- and adenoviruses, MST markers (GenBac3 and HF183), E. coli, coliform bacteria, C. perfringens and protozoans. Moreover, a sewage sample from the municipal wastewater treatment plant was collected on 26th October 2016 and analyzed for sapoviruses and protozoans. 2.1.2. Outbreak II In outbreak II, both a drinking water pipe and a wastewater pipe were broken at the same site. It was suggested that the drinking water pipe had leaked for several months near the wastewater pipe, and eventually, this caused a collapse of the waterlogged soil and the breakage of the sewer. The under pressure event in the drinking water distribution system during the search of the leakage on 22nd January 2018 most probably caused the inflow of wastewater from the contamination site into the drinking water network. The contamination site with broken pipes was detected eight days after the assumed contamination event on 30th January 2018. Drinking water originating from the groundwater source was alkalized and UV-treated prior to distribution. Drinking water was not chlorinated. Mitigation actions during the outbreak included the boil water advisory for four weeks (from 27th January to 23rd February 2018) and chlorination for six weeks (from 27th January to 10th March 2018) with chlorine levels ranging from <1 mg/L–2 mg/L and including 3–5 days intensive chlorination with chlorine levels 5 mg/L–10 mg/L (started on 6th February 2018). In addition, alternative water sources were arranged for six weeks (from 29th January to 11th March 2018). The drinking water contamination affected approximately 4000 people. During January–February, 463 persons with gastrointestinal illness contacted local primary health care. Some cases occurred already before the assumed under pressure event, but most of the patient cases appeared from 24th January–30th January 2018 (Figure 2b). Symptoms lasted on average for two days and included diarrhea (76%, 352/463), vomiting (65%, 299/463) and fever (32%, 150/463). Stool samples were collected during the acute phase of the outbreak and were analyzed in clinical laboratories with routine tests for enteric viruses, pathogenic bacteria, and protozoans (Table 1). Since the clinical laboratory method did not distinguish between norovirus genogroups, seven samples were further analyzed by the genogroup-specific real-time RT-PCR [32]. FIB (E. coli, coliform bacteria, intestinal enterococci, and C. perfringens) were analyzed from drinking water samples collected between 27th January 2018 and 5th March 2018 in a local environmental laboratory. Drinking water samples for pathogen (sapo-, noro- and adenovirus, pathogenic E. coli strains, Campylobacter spp., Giardia spp. and Cryptosporidium spp.) and MST marker (GenBac3 and HF183) analyses were taken before chlorination on 27th January 2018 (n = 1), after the initial low level chlorination (<1 mg/L) on 29th January 2018 and 6th February 2018 (n = 2) and after the intensive chlorination on 14th February 2018 (n = 3, DEUF). A surface water sample from the contamination site and a biofilm sample from water meter were collected on 31st January and were analyzed for the selected microbes (Table 2). 2.2. Environmental Investigation 2.2.1. Sample Collection and Concentration Water was collected into sampling bottles or large volume (100–200 L) samples were taken using dead-end ultrafiltration (DEUF) method [33]. After the water sampling, sodium thiosulphate was used to inactivate the chlorine from the samples during the transport prior to microbiological analyses. In the DEUF method, water samples were collected using ASAHI Rexeed-25A (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) ultrafilters with an average flow rate of 3 L/min. Backflush of the ultrafilters was performed with 500 mL of backflush solution (0.5% Tween 80, 0.01% sodium polyphosphate and 0.001% Y-30 antifoam emulsion). The secondary concentration of DEUF eluates was performed by filtration through Millipore Express PLUS membrane filters (outbreak I, pore size 0.22 µm, Merck KGaA, Darmstadt, Germany) or Nuclepore polycarbonate (PC) filters (outbreak II, pore size 0.4 µm, Whatman, Kent, UK) and/or polyethylene glycol (PEG) precipitation (Table S1). In PEG precipitation, the sample (pH 7–7.5) was mixed with 1% BSA (only for drinking water samples), 0.9 M NaCl and 12% PEG8000 and kept for at least 2 h at 4 ◦C. After incubation, the sample was centrifuged 10 000× g for 30 min at 4 ◦C and the pellet was suspended in PBS. Biofilm from water meters was detached and collected as previously described [34]. Before further analyses, all biofilm samples were sonicated for 1 min in 40 kHz (Branson Ultrasonics, Danbury, USA). Biofilm samples were concentrated by filtration through PC filters and PEG precipitation of the filtrate. 2.2.2. Detection of Enteric Virus Genomes Enteric viruses were analyzed in raw water and drinking water samples either with low volume (1–2 L) adsorption-elution methods or a large volume DEUF method. Low volume samples were concentrated using disc filters (Sartolon polyamide, Sartorius, Göttingen, Germany; Zetapor, Amf-Cuno, Meriden, USA or Nanoceram, Argonide, Sanford, USA) as previously described [35] or modified from Maunula et al. [36], Schultz et al. [37] and Kim and Ko [38]. Samples from contamination sites were analyzed from a volume of 400 mL by PEG precipitation (outbreak I) or extracted directly from a volume of 2.5 mL (outbreak II). Viral RNA and DNA were extracted from the low volume concentrates using the High Pure Viral RNA Kit and High Pure Viral Nucleic Acid Kit (Roche Diagnostics GmbH, Mannheim, Germany), respectively, or the Nuclisens Minimag system (bioMerieux, Marcy-l’Etoile, France). In addition, High Pure Viral Nucleic Acid Large Volume Kit (Roche Diagnostics GmbH) was used with PEG precipitates and directly extracted samples. A sewage sample from municipal wastewater treatment plant was directly treated with Nuclisens kit and the nucleic acid was further purified using OneStep™ PCR Inhibitor Removal (Zymo Research, Irvine, USA). Extractions were made according to the manufacturers’ instructions. Extracted nucleic acids were stored at −75 ◦C. For noroviruses, the real-time RT-qPCR assays were carried out in one step, separately for genogroups I and II, using the TaqManfiFast Virus 1-Step Master Mix (Thermo Fisher Scientific, Austin, TX, USA) as well as primers and probes as previously described [35,39]. For sapoviruses, the real-time RT-qPCR assays were carried using the same protocol with noroviruses [39] or using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) with a slightly modified norovirus protocol [40]. Sapovirus primers and probes were according to the study by Oka et al. [41] or van Maarseveen et al. [42]. Adenoviruses were detected using primers and a probe described by Jothikumar et al. [43] with the real-time qPCR assay as described previously [44]. The adenovirus real-time qPCR program was 95 ◦C for 10 min, followed by 45 cycles at 95 ◦C for 15 s and 60 ◦C for 1 min. The virus assays were carried out using the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, USA) or the RotorGene PCR cycler (Qiagen). Quantification of genome copies (GC) of each virus was done using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies, Leuven, Belgium). The standard curves were included in each run. The quality of virus extraction was controlled by positive and negative process controls through all stages of the analytical steps. Spiked mengovirus strain VMC0 and human adenovirus 40 (ATCC VR-931) were used as a positive process controls and sterile deionized water as a negative process control. External amplification controls (EACs) were used to control norovirus GI and GII RT-PCR inhibition in samples as previously described [35]. No inhibition was detected in samples tested negative in norovirus analysis. Genotyping of sapovirus and norovirus was performed with conventional RT-PCR using One-Step RT-PCR kit (Qiagen). Sapovirus genome was amplified using primers p289 and p290 [45]. Norovirus RNA was amplified in polymerase region A according to Vinjé et al. [46]. The amplicons visualized in gel electrophoresis were sent to Sanger sequencing in the Institute of Biotechnology. Sequences were assigned using the Norovirus Genotyping Tool [47] or with NCBI database using BLAST (basic local alignment search tool). 2.2.3. Enumeration of Indicator Bacteria Standard methods were used to enumerate E. coli, coliform bacteria, intestinal enterococci, and C. perfringens count from water and biofilm samples. In brief, E. coli and coliform bacteria were analyzed using membrane filtration with LES Endo medium [48] and Chromocult Coliform Agar medium [49] or by using the most probable number (MPN) method based on Colilert-18 QuantiTray [50]. The counts of intestinal enterococci were analyzed using the membrane filtration on Slanetz and Bartley medium [51] or Enterolert (IDEXX Laboratories Inc, Westbrook, USA). Vegetative cells and spores of C. perfringens were enumerated on tryptose sulfite cycloserine agar following the international standard [52]. 2.2.4. Detection of Microbial Source Tracking (MST) Markers MST markers were analyzed from nucleic acids extracted from samples of raw water, drinking water and biofilms of water meters either using DEUF method or PC filters. Samples from contamination sites were extracted directly. The nucleic acids were extracted using Chemagic DNA Plant kit (Perkin Elmer, Waltham, USA). Complementary DNA was synthesized as previously described (outbreak I) [34] or by using Superscript IV VILO (outbreak II, Thermo Fisher Scientific, Waltham, USA). MST markers (GenBac3 and HF183) were quantified using DNA-based qPCR assays and RNA-based RT-qPCR assays as described earlier by Pitkänen et al. [27]. The assays were carried out with the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using standard curves generated with serially diluted gBlocksfiGene Fragments (Integrated DNA Technologies). 2.2.5. Detection of Bacterial Pathogens The presence/absence of thermotolerant Campylobacter spp. was determined using culture-based selective enrichment methods following the principles of the international standard [53]. Pathogenic E. coli strains (ETEC, EPEC, EHEC, and EAEC) were analyzed from nucleic acid aliquots with in-house PCR method in a clinical laboratory [54]. 2.2.6. Detection of Protozoans In outbreak I, the aliquots of nucleic acids extracted with the Nuclisens Minimag system or Chemagic DNA Plant kit were sent to protozoan (G. lamblia, E. histolytica, Cryptosporidium spp. and D. fragilis) analysis in the UnitedMedix Laboratories Ltd. In outbreak II, Giardia spp. and Cryptosporidium spp. were analyzed from drinking water with qPCR using primers and probes described in Hill et al. [55] and Jothikumar et al. [56], respectively, from nucleic acid subsamples. Samples from the contamination site and water meter biofilms were analyzed using the immunomagnetic separation method (IMS) based on standard ISO 15553 [57]. In brief, the sample was centrifuged (15 min, 1100 g) and IMS (Dynabeads G/C Combo, IDEXX laboratories Inc) was done for pellet in the volume of 10 ml. Samples were stained with FITC and DAPI (EasyStain, bioMerieux) and analyzed with epifluorescence microscopy. 3. Results 3.1. Clinical Findings Sapoviruses were found from patients’ stool samples in both outbreaks (Table 1). In outbreak II, sapovirus GIV was detected in one patient sample subjected for sequencing. Moreover, adenoviruses were detected in outbreak I and noroviruses and astroviruses in outbreak II. Noroviruses were not detected in outbreak I. In outbreak II, noroviruses were detected more frequently than sapoviruses. Twelve of the sixteen norovirus positive samples were sequenced successfully and identified as genotypes GI.P7 (n = 11) and GI.P6 (n = 1). In addition, seven out of 16 norovirus positive samples were further analyzed by the genogroup-specific real-time RT-PCR. Of these samples, norovirus GI was detected in all seven samples and norovirus GII in one of seven samples. Sporadic bacterial infections (outbreak II) and D. fragilis (outbreak I) were also found in patient samples. 3.2. Environmental Investigations In outbreak I, only E. coli and coliform bacteria were analyzed before the start of the chlorination and were detected in one of the two water samples (Table 2). In addition, low counts of coliform bacteria were detected in three out of 91 water samples taken after chlorination on 17th October 2016 and 19th October 2016, and two out of nine biofilm samples on 28th November 2016. Water samples were collected for pathogen and MST-marker analyses only after chlorination. Traces of GenBac3 rRNA were found from one of the three samples on 28th November 2016. In the sample taken from the contamination site, high numbers of both pathogens and indicators were detected. Typing of sapovirus was unsuccessful for contamination site sample. The raw water sample was positive only for GenBac3 rRNA and coliform bacteria. Sewage sample taken from the municipal wastewater treatment plant on 26th October 2016 was positive for sapovirus (genotype GI.2, accession number MK689409) and D. fragilis. In outbreak II, samples were taken before and after the start of the chlorination. Low E. coli and intestinal enterococci counts as well as both MST markers (GenBac3 and HF183) were detected from the water before chlorination (Table 3). In a sample taken after the start of the chlorination, sapovirus and genes of enteropathogenic E. coli (EPEC) were detected from the drinking water. Sapovirus genotyping was attempted but failed most probably due to the small number of viruses in the sample. Findings of fecal microbes in drinking water, however, led to the decision to perform intensive chlorination. After intensive chlorination, intestinal enterococci were detected in two out of 48 water samples taken from the same site on 15th February and 22nd February 2018. Also, small numbers of GenBac3 rDNA and rRNA copies were detected in three water samples on 15th February 2018. The sample taken from the contamination site on 31st January 2018 contained the same pathogens than detected from the patients and high levels of fecal indicators. Typing of sapovirus was unsuccessful for the contamination site sample. A biofilm sample from water meter on 31st January 2018 was positive only for GenBac3 rDNA and rRNA. 4. Discussion This study presents two waterborne outbreaks caused by drinking water pipe breakage and subsequent contamination of the distribution network. The sudden onset of symptoms and clinical picture of the illness fitted symptoms of viral infection [12]. Stool samples collected from patients confirmed that most of the clinical cases were due to enteric virus infections and sapoviruses were found from patients’ samples in both outbreaks. Sapovirus genotype GI.2 was detected from a sewage sample in outbreak I and sapovirus GIV in a one patient sample in outbreak II. Genotype GI.2 is one of the predominant genotypes worldwide and sapovirus GIV predominated in several countries in 2007 [12]. Unfortunately, patient samples were not sequenced more comprehensively to determine sapovirus genotypes. In many countries, including Finland, norovirus has been the most common causative agent in waterborne outbreaks [5,6], while the linkage of sapovirus infections to possible waterborne spread and outbreaks is rare [9,10]. To our knowledge, this is the first outbreak study worldwide describing the detection of sapovirus in drinking water. In the future, the significance of this emerging virus may increase and thus testing for sapovirus is important to include in waterborne outbreak investigations. In both outbreaks, untreated municipal wastewater entered into the drinking water distribution network. Raw wastewater reflects the infection burden among the population and can contain a wide variety of pathogens. Water samples taken from the contamination sites contained the same pathogens that were detected from patient samples. However, of these pathogens only sapovirus and EPEC were detected in drinking water in outbreak II. In outbreak I, no water samples were obtained for pathogen analyses before start of the chlorination, which is presumably the main reason behind the non-detection of pathogens from drinking water. However, the first samples taken before chlorination in outbreak I were positive for coliform bacteria indicating the deficiency in the water quality. In outbreak investigations, it is important to collect enough water before mitigation actions for possible future use, in this case e.g., for sapovirus analysis. However, the pathogen sampling should not delay the actions necessary to prevent further spread of infections. Overall, pathogens are not analyzed as comprehensively as fecal indicator bacteria (E. coli and intestinal enterococci) in environmental investigations of outbreaks. This is partly due to their higher cost compared to indicator analyses and the need for expert laboratories to conduct the tests. Even though FIB has often been insufficient to prove the safety of water [58–61], in this study, these indicators were able to detect the water contamination in both outbreaks. In outbreak I, coliform bacteria and in outbreak II, coliform bacteria, E. coli, intestinal enterococci, and C. perfringens were detected in drinking water. Noteworthy, sporadic findings of intestinal enterococci were detected in water even after intensive chlorination in outbreak II. These findings support the use of traditional FIB in water quality assessments during outbreak investigation. However, the value of indicators in the prediction of water contamination seems to be case-specific and may require massive contamination as was the situation in the outbreaks described herein and in previous outbreaks described by Kauppinen et al. [35]. In this study, the suitability of molecular qPCR assays for fecal source tracking markers (HF183 and GenBac3), along with the traditional FIB was evaluated during waterborne outbreak investigations. The use of genetic source identifiers may provide more sensitive detection of the contamination especially when the assays are targeted to rRNA transcripts in addition to the rDNA [27]. Further, by using a host-specific marker, such as HF183 it is possible to identify the source of the contamination. In contamination sites, HF183 and GenBac3 numbers were comparable or higher than the numbers of pathogenic viruses. Moreover, the markers targeting to host-specific sequences from Bacteroidales clearly outnumbered traditional FIB in contamination site samples and thus could be considered for use as specific and sensitive fecal indicators of drinking water quality. Particularly, the human-specific marker HF183 showed promising results and the findings in water were in concordance with pathogen findings. On the other hand, GenBac3 prove to be a very sensitive marker and small GenBac3 copy numbers were found in drinking water after chlorination in both outbreaks and even after intensive chlorination in outbreak II. Interestingly, Diston et al. [62] found in a Swiss groundwater study that genetic markers of Bacteroidales are sensitive indicators, but due to the higher presence of these markers compared to enteric viruses may overestimate the risk from enteric viral pathogens. Thus, more data is needed for the correct interpretation of the significance of GenBac3 marker detection after intensive chlorination in terms of health risk assessment. Mitigation actions, including boil water advisory, providing an alternative drinking water source and chlorination of the drinking water network, were conducted in both outbreaks and proved efficient in controlling the outbreaks. Previous studies have shown the long persistence of enteric viruses and protozoans in drinking water distribution systems in cases without proper treatment or removal of the contamination source [35,63–65]. Even though chlorine has been shown to be an efficient decontaminant in the drinking water distribution system [65], the possible stagnant locations (i.e., dead-ends) in the network and deposits accumulated on the inner surfaces of the old pipes may hamper the success of the chlorination. These factors may explain the sporadic microbial findings in water samples followed chlorination. Therefore, it is important to allow sufficient time for chlorination and to ascertain the purity of the water with microbiological analyses as was carried out in these outbreaks. The aging water infrastructure [66] and improper drinking water pipeline construction practices pose a major challenge for water supply and may compromise drinking water safety even more often in the future. D. fragilis detection from patient samples induced media headlines and health concerns among the water consumers. The questionable pathogenesis of this parasite [67,68] initiated a more throughout epidemiological investigation (unpublished results). Lack of knowledge related to the drinking waterborne transmission of D. fragilis increased the uncertainty of crowds and up kept the media attention on the topic over a prolonged time. 5. Conclusions To our knowledge, this is the first outbreak study describing the detection of sapovirus in drinking water. Further, herein we proved the suitability of source tracking identifiers to be applied in waterborne outbreak investigation along with pathogens and water quality indicator analyses. Main conclusions are as follows: • This study highlights the importance of sapovirus as a waterborne pathogen, and warrants the need for testing of multiple pathogens during outbreak investigation • The MST markers proved useful in the detection of contamination and especially HF183 findings were in concordance with the pathogen results, supporting its use in drinking water outbreak investigations • Boil water advisory, alternative drinking water source and chlorination were effective mitigation actions during the outbreaks • The role of D. fragilis as human pathogen and its drinking waterborne transmission potential requires further studies

What were the infrastructure complaints?

1493

Value of syndromic surveillance in monitoring a focal waterborne outbreak due to an unusual Cryptosporidium genotype in Northamptonshire

The United Kingdom (UK) has several national syndromic surveillance systems. The Health Protection Agency (HPA)/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data from a national telephone helpline, while the HPA/ QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems. Data from both of these systems were used to monitor a local outbreak of cryptosporidiosis that occurred following Cryptosporidium oocyst contamination of drinking water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, in June 2008. There was a peak in the number of calls to NHS Direct concerning diarrhoea that coincided with the incident. QSurveillance data for the local areas affected by the outbreak showed a significant increase in GP consultations for diarrhoea and gastroenteritis in the week of the incident but there was no increase in consultations for vomiting. A total of 33 clinical cases of cryptosporidiosis were identified in the outbreak investigation, of which 23 were confirmed as infected with the outbreak strain. However, QSurveillance data suggest that there were an estimated 422 excess diarrhoea cases during the outbreak, an increase of about 25% over baseline weekly levels. To our knowledge, this is the first time that data from a syndromic surveillance system, the HPA/QSurveillance national surveillance system, have been able to show the extent of such a small outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide data at local health district (primary care trust) level. The Cryptosporidium contamination incident described demonstrates the potential usefulness of this information, as it is unusual for syndromic surveillance systems to be able to help monitor such a small-scale outbreak. Introduction As syndromic surveillance systems usually capture data already collected for other purposes, and monitor generic symptoms and/or clinically diagnosed disease, they provide information at an earlier stage of illness (compared with laboratory-confirmed diagnoses), so that action can be taken in time to substantially reduce the impact of disease. Some systems, for example, the Royal College of General Practitioners Weekly Returns Service, are now well established, with many years of historical data that allow monitoring of longer-term disease trends [1]. They have the ability to provide early warning of, for example, seasonal rises in influenza and can trigger public health action, such as a recommendation to prescribe antiviral drugs in line with national guidance [2-4]. They can also provide reassurance to incident response teams and the general public that an incident has not caused adverse health effects – for example, following an explosion at the Buncefield oil storage depot in Hemel Hempstead, United Kingdom (UK), in 2005, syndromic surveillance confirmed that there were no unusual rises in community-based morbidity linked to the incident [5]; following the eruption of the Eyjafjallajökull volcano in Iceland in April 2010 similar assurance was given about lack of impact on community morbidity [6]. Health departments are increasingly expected to monitor health effects of natural events such as heat wave or flooding, or implement surveillance – of which syndromic surveillance plays a major role – for mass gatherings such as the Olympics or football World Cup [7-9]. Systems in France, Australia and Taiwan use data from emergency departments [10-12], a Canadian system uses over-the-counter pharmacy sales [13,14], and in the Netherlands data from both syndromic and surrogate data sources, such as employee absence records and prescription medications dispensed by pharmacies, are included in surveillance systems [15,16]. Currently systems based on Internet searches via search engines or on queries submitted to medical websites are being developed [17,18]. In the UK, the HPA/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data collected from the NHS Direct telephone helpline [19], while the HPA/QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems [20]. The HPA Real-time Syndromic Surveillance Team is a small team that coordinates a number of syndromic surveillance systems within the HPA and takes a lead for syndromic surveillance in England [21]. This paper describes the support provided by the team to the local incident management team during a local cryptosporidiosis outbreak and shows the use of syndromic surveillance in monitoring the extent of an outbreak using the HPA/NHS Direct and HPA/QSurveillance national surveillance systems. Cryptosporidiosis Cryptosporidium is a protozoan parasite that can cause an infection in people, cattle and sometimes other animals [22]. Cryptosporidiosis is most common in children aged between one and five years, but it can affect all ages. Those with impaired immune systems are likely to be most seriously affected. Symptoms usually appear between three and 12 days after initial exposure and include watery diarrhoea, stomach pains, dehydration and fever. In its transmissible form, called an oocyst, the parasite is protected by an outer shell, which allows it to survive in the environment for a long time. Transmission occurs most often via the faeco-oral route through person-to-person or animalto-person contact, but people may also be infected by consuming contaminated water or food or by swimming in contaminated water. Although uncommon, the largest outbreaks have occurred following contamination of drinking water [23,24]. Normal chlorine disinfection procedures do not kill the oocysts, so they are removed by filtration and water companies carry out routine monitoring of treated water. Description of the incident On 25 June 2008 the local Health Protection Unit was informed by Anglian Water of an exceedence in the level of Cryptosporidium oocysts found in water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, during 19 to 24 June 2008 [25]. The reservoir supplied a population of more than 250,000 in the Northampton area. A notice advising people in the affected areas to boil all drinking water was issued on 25 June 2008 and public health messages were circulated to local health services and to the general public via the media. Those members of the public who were concerned about health risks associated with the incident were asked to ring NHS Direct for clinical advice [26]. The HPA wrote to local GPs and hospitals asking them to monitor potential patients for signs and symptoms of Cryptosporidium infection and to submit faecal specimens to the local hospital diagnostic laboratory if patients presented with diarrhoea. Samples from 34 patients where Cryptosporidium infection was identified were sent to the UK Cryptosporidium reference unit for typing. On 30 June 2008, the Cryptosporidium oocysts found in the reservoir water were confirmed as being of the rabbit genotype Cryptosporidium cuniculus [27]. Subsequently, a dead rabbit was found in a treated water tank at the water treatment works. The genotype of Cryptosporidium oocysts in the rabbit’s large bowel was indistinguishable from that of the oocysts found in the water [27]. After remediation of the water supply and distribution, the ‘boil water notice’ was lifted on 4 July and the following day the first case of cryptosporidiosis linked to the incident was identified by the reference laboratory (this case was infected with C. cuniculus). During the course of the outbreak (24 June – 18 July 2008, the dates of symptom onset in the first and last case, respectively), 23 cases of cryptosporidiosis were confirmed as being infected with C. cuniculus; one of the 23 was a secondary case. The HPA Real-time Syndromic Surveillance Team provided data in order to aid the response to this incident and the first syndromic surveillance report was circulated to the incident management team and other relevant people in the HPA on 27 June 2008. Data from the HPA/NHS Direct and HPA/QSurveillance systems were provided in a series of regular reports, initially daily and eventually weekly, until the final report on 21 August 2008. Each report included a summary interpretation and more detailed data on diarrhoea, gastroenteritis and vomiting indicators. Methods Surveillance systems HPA/NHS Direct surveillance system NHS Direct is a 24-hour nurse-led telephone helpline that provides health information and advice to the general public. Nurses use a computerised clinical decision support system – the NHS Clinical Assessment System (NHS CAS) – to handle calls. This assessment system uses approximately 200 computerised symptom-based clinical algorithms. Nurses assign the call to the most appropriate algorithm and the patient’s symptoms determine the questions asked and the action to be taken following the call (call outcome), which could be guidance on self-care or they could be referred to their GP or advised to attend a hospital emergency department. No attempt is made to provide a formal diagnosis. Daily NHS Direct data are received by the Real-time Syndromic Surveillance Team, where the number and type of calls received during the previous day are analysed and interpreted. Call proportions are calculated by age group and algorithm against the total number of calls received. HPA/QSurveillance system The HPA/QSurveillance national surveillance system was set up by the University of Nottingham, United Kingdom, and Egton Medical Information Systems (EMIS), a supplier of general practice computer systems, in collaboration with the HPA. It comprises a network of more than 3,500 general practices throughout the UK, covering more than 22 million patients (about 38% of the population [28]). Aggregated data on GP consultations for a range of indicators are automatically uploaded daily from GP practice systems to a central database. Data are routinely reported on a weekly basis; however, daily reporting is possible for specific incidents. Reports are provided at national or regional level (strategic health authority, SHA) and by local health district (primary care trust, PCT). Analysis of surveillance data NHS Direct call proportions for gastrointestinal syndromes (diarrhoea and vomiting) for the East Midlands region in England, where Northampton is situated, were examined during the outbreak (24 June – 18 July 2008) and compared with those for England and Wales. A series of control charts for diarrhoea calls are routinely used to monitor significant rises in the numbers of calls received. Control charts are calculated by assuming that calls follow a Poisson distribution with the total number of calls as an offset: a model is fitted to each region and symptom separately [29]. The model takes into account call variation caused by weekends, public holidays and the time of year – variables that can affect the number of calls received by NHS Direct. A value above the upper limit of the 99.5% confidence interval would be considered to be unusual. The seven-day moving average for diarrhoea calls was also monitored. The number and percentage of calls for diarrhoea in the East Midlands region were presented by call outcome and the number of calls in the Northampton (NN) postcode districts and in particular the number of calls in the NN11 and NN12 postcode districts, which were most affected by the incident. QSurveillance national consultation rates per 100,000 population for diarrhoea (in the age groups under five years, five years and over, and all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with rates for the same period in 2007 (data not presented). Consultation rates by region for 2008 for diarrhoea (all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with those for the East Midlands region. The gastroenteritis indicator includes all cases of diarrhoea and/or vomiting. Consultation rates and standardised incidence ratios (SIRs) – calculated using the UK as the standard population – for diarrhoea, gastroenteritis and vomiting were compared for the UK, Yorkshire and Humberside, East Midlands, Leicestershire, Northamptonshire and Rutland SHA, and Daventry and South Northants PCT, Northamptonshire Heartlands PCT and Northampton PCT. Yorkshire and Humberside was not an affected region but was included as a control. The area supplied by the Pitsford Reservoir included the three PCTs, which were all within the Leicestershire, Northamptonshire and Rutland SHA. The consultation rates and SIRs were compared for the period from week 16 to week 35 of 2008 in order to compare the rates before and after the Cryptosporidium exceedance, which took place in week 26. Estimates of excess numbers of cases of diarrhoea occurring during and following the Cryptosporidium outbreak were made by calculating the mean consultation rate over a period of five weeks before and after the incident (weeks 20–24 and weeks 31–35, respectively). For each of the three PCTs, the calculated mean rate was applied to the PCT population to estimate the number of cases that would be expected each week. The actual consultation rates for diarrhoea for weeks 25 to 30 were used to estimate the number of cases for the PCT population each week. The expected number of cases was subtracted from the estimated number of cases in the PCT population to give the estimated number of excess cases. Results HPA/NHS Direct surveillance system A peak in the number of calls for diarrhoea in the East Midlands was recorded in 25–26 June 2008, the period that coincided with the contamination incident and the associated media coverage (Figure 1). The neighbouring areas of the West Midlands, Yorkshire and the Humber, and East of England showed no increase in the number of calls for diarrhoea. The peak produced a control chart exceedance for calls for diarrhoea on 25 June 2008 (Figure 2), when the proportion of calls exceeded the upper limit of the 99.5% confidence interval. There were further confidence interval exceedances on 26 and 28 June (which were not control chart exceedances). There was no peak in calls for vomiting or control chart exceedance for these calls in the East Midlands. HPA/QSurveillance national surveillance system The East Midlands region had significantly high consultation rates for diarrhoea and gastroenteritis in week 25 (16–22 June), week 26 (23–29 June 2008, when the contamination incident was reported) and in the following four weeks. Within this region. Leicestershire, Northamptonshire and Rutland SHA had slightly raised consultation rates and significant SIRs across weeks 25 to 30 that were not seen in the neighbouring Trent SHA. At PCT level, all three of the PCTs in the area affected by the incident showed increased consultation rates for diarrhoea (Table 1) and gastroenteritis (Table 2) with SIRs significantly above the UK rate in week 26. Daventry and South Northants PCT also had a raised SIR for both indicators in week 25, and although Northamptonshire Heartlands and Northampton PCTs did not have SIRs significantly above that of the UK in week 25, the rise in consultation rates for diarrhoea and gastroenteritis began during week 25. In Northampton PCT, consultations for both diarrhoea and gastroenteritis peaked in the week following the contamination incident, week 27, returning to normal levels by week 30 (Figure 3A and 3B). A similar effect can be seen in Northampton Heartlands PCT. Daventry and South Northants PCT also showed an increase, but appeared to have consistently higher rates for both indicators. This was the area with the smallest population so the rates were more variable than in the other PCTs and we therefore interpreted these results with caution. The consultation rates for vomiting during weeks 25 to 30 in the East Midlands were not unusual at SHA or PCT level (data not presented). Discussion We have demonstrated the sensitivity of syndromic surveillance in detecting this small Cryptosporidium outbreak and the value of the surveillance in being able to describe the extent of its spread. Both the HPA/NHS Direct and HPA/QSurveillance systems showed demonstrable increases in calls and consultations for diarrhoea that were linked to the outbreak. QSurveillance consultations appeared to increase across the PCTs immediately affected but not in the surrounding area. Both the HPA/NHS Direct and HPA/QSurveillance systems showed a clear signal at the time of the incident and we were able to describe the extent of the impact on pre-primary care and primary care services. The HPA/QSurveillance system showed a rise in consultation rates for gastrointestinal symptoms that began the week before the outbreak, consistent with the period when Cryptosporidium was present in the water leaving the Pitsford Reservoir (19–24 June 2008) and with the onset of symptoms in the first outbreak case on 24 June. Although only 33 cases were identified by the outbreak investigation team, of which 23 were confirmed as having the outbreak Cryptosporidium strain, our syndromic surveillance data detected this limited outbreak. Data also suggested a more widespread increase in general gastrointestinal symptoms around the time of the outbreak, with an estimated 422 excess diarrhoea cases; these excess cases represented an increase of about 25% above normally expected levels. It is highly probable that a proportion of these excess cases may have resulted from the increased publicity surrounding the incident – for example, it is likely that media reports contributed to the large peak in calls detected by the HPA/NHS Direct surveillance system on the day the boil water notice was issued, and could also have impacted on the GP consultation rate. It has been previously shown that reporting of mumps cases is sensitive to media coverage, with a rise in clinically reported cases following newspaper reports [30]. A similar mechanism could account for some of the excess GP consultations as cases experiencing gastrointestinal symptoms may have been more likely to consult their GP, whereas in normal circumstances they would have cared for themselves at home. It is also possible that the surveillance shows outbreak-associated cases that did not come to the attention of the outbreak team, perhaps because symptoms were not sufficiently severe to warrant further investigation, or stool samples were not provided for testing. It is interesting to note that there was no demonstrable impact on the number of calls for vomiting (which is not a prominent clinical feature of cryptosporidiosis). Other common community-based pathogens such as norovirus and rotavirus were at low levels, as is normal for that time of year [31]. In this instance, public health authorities had already been alerted to a potential problem by the water company, although the extent of the outbreak was detected by syndromic surveillance. In 2003 the syndromic surveillance systems in the city of New York, United States, were able to detect an increase in diarrhoeal illness following a power outage when there was no other indication of citywide illness [32]. The New York system covers a population of nine million, but does not regularly detect localised outbreaks [33]. It has been shown previously that the HPA/NHS Direct surveillance system would be unlikely to detect a Cryptosporidium outbreak unless call volumes are high (72% chance of detection if nine-tenths of cases called NHS Direct) [29], although the value of syndromic surveillance for such outbreaks has been recognised [34]. The system detected the East Midlands Cryptosporidium outbreak that affected a smaller population than that covered by the New York system. The three PCTs affected have a combined population of around 600,000, of which just over half use GP practices reporting to QSurveillance, yet this syndromic surveillance system was able to describe an increase in consultation rates for diarrhoea and gastroenteritis around the time of the outbreak. Limitations of the data There was extensive media reporting of the incident that may have affected both the HPA/NHS Direct and HPA/QSurveillance systems and contributed to the increase in reported gastrointestinal symptoms around the time of the contamination incident. However, the rise in consultation rates for diarrhoea began before the outbreak had been detected and therefore cannot be attributed to media coverage. The HPA/NHS Direct and HPA/QSurveillance systems monitor general symptoms and so could only monitor the relevant symptoms of diarrhoea and vomiting. They are not able to detect Cryptosporidium cases, as this would require laboratory confirmation of diagnosis, so some of the estimated excess cases could be unconnected with this incident. This outbreak was discovered by other means but both the HPA/NHS Direct and HPA/ QSurveillance systems were able to describe the extent of the disease in the general population and provide reassurance that there was no widespread impact. Compared with other populations, older people and ethnic minorities are less likely to call NHS Direct [29], and although this should not prevent detection of gastrointestinal symptoms as a result of drinking water contamination as this would affect the whole population, this may reduce the signal from the system [35]. With such large surveillance systems, there will be ‘background noise’ in the data, so procedures must be in place to correctly interpret the data and set appropriate thresholds for action. Conclusion To our knowledge, this is the first time that PCT-level data from a syndromic surveillance system, the HPA/ QSurveillance national surveillance system, have been able to show the extent of such a limited outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide PCT-level data and this Cryptosporidium contamination incident demonstrates the potential usefulness of this system.

What happened?

1494

Value of syndromic surveillance in monitoring a focal waterborne outbreak due to an unusual Cryptosporidium genotype in Northamptonshire

The United Kingdom (UK) has several national syndromic surveillance systems. The Health Protection Agency (HPA)/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data from a national telephone helpline, while the HPA/ QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems. Data from both of these systems were used to monitor a local outbreak of cryptosporidiosis that occurred following Cryptosporidium oocyst contamination of drinking water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, in June 2008. There was a peak in the number of calls to NHS Direct concerning diarrhoea that coincided with the incident. QSurveillance data for the local areas affected by the outbreak showed a significant increase in GP consultations for diarrhoea and gastroenteritis in the week of the incident but there was no increase in consultations for vomiting. A total of 33 clinical cases of cryptosporidiosis were identified in the outbreak investigation, of which 23 were confirmed as infected with the outbreak strain. However, QSurveillance data suggest that there were an estimated 422 excess diarrhoea cases during the outbreak, an increase of about 25% over baseline weekly levels. To our knowledge, this is the first time that data from a syndromic surveillance system, the HPA/QSurveillance national surveillance system, have been able to show the extent of such a small outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide data at local health district (primary care trust) level. The Cryptosporidium contamination incident described demonstrates the potential usefulness of this information, as it is unusual for syndromic surveillance systems to be able to help monitor such a small-scale outbreak. Introduction As syndromic surveillance systems usually capture data already collected for other purposes, and monitor generic symptoms and/or clinically diagnosed disease, they provide information at an earlier stage of illness (compared with laboratory-confirmed diagnoses), so that action can be taken in time to substantially reduce the impact of disease. Some systems, for example, the Royal College of General Practitioners Weekly Returns Service, are now well established, with many years of historical data that allow monitoring of longer-term disease trends [1]. They have the ability to provide early warning of, for example, seasonal rises in influenza and can trigger public health action, such as a recommendation to prescribe antiviral drugs in line with national guidance [2-4]. They can also provide reassurance to incident response teams and the general public that an incident has not caused adverse health effects – for example, following an explosion at the Buncefield oil storage depot in Hemel Hempstead, United Kingdom (UK), in 2005, syndromic surveillance confirmed that there were no unusual rises in community-based morbidity linked to the incident [5]; following the eruption of the Eyjafjallajökull volcano in Iceland in April 2010 similar assurance was given about lack of impact on community morbidity [6]. Health departments are increasingly expected to monitor health effects of natural events such as heat wave or flooding, or implement surveillance – of which syndromic surveillance plays a major role – for mass gatherings such as the Olympics or football World Cup [7-9]. Systems in France, Australia and Taiwan use data from emergency departments [10-12], a Canadian system uses over-the-counter pharmacy sales [13,14], and in the Netherlands data from both syndromic and surrogate data sources, such as employee absence records and prescription medications dispensed by pharmacies, are included in surveillance systems [15,16]. Currently systems based on Internet searches via search engines or on queries submitted to medical websites are being developed [17,18]. In the UK, the HPA/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data collected from the NHS Direct telephone helpline [19], while the HPA/QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems [20]. The HPA Real-time Syndromic Surveillance Team is a small team that coordinates a number of syndromic surveillance systems within the HPA and takes a lead for syndromic surveillance in England [21]. This paper describes the support provided by the team to the local incident management team during a local cryptosporidiosis outbreak and shows the use of syndromic surveillance in monitoring the extent of an outbreak using the HPA/NHS Direct and HPA/QSurveillance national surveillance systems. Cryptosporidiosis Cryptosporidium is a protozoan parasite that can cause an infection in people, cattle and sometimes other animals [22]. Cryptosporidiosis is most common in children aged between one and five years, but it can affect all ages. Those with impaired immune systems are likely to be most seriously affected. Symptoms usually appear between three and 12 days after initial exposure and include watery diarrhoea, stomach pains, dehydration and fever. In its transmissible form, called an oocyst, the parasite is protected by an outer shell, which allows it to survive in the environment for a long time. Transmission occurs most often via the faeco-oral route through person-to-person or animalto-person contact, but people may also be infected by consuming contaminated water or food or by swimming in contaminated water. Although uncommon, the largest outbreaks have occurred following contamination of drinking water [23,24]. Normal chlorine disinfection procedures do not kill the oocysts, so they are removed by filtration and water companies carry out routine monitoring of treated water. Description of the incident On 25 June 2008 the local Health Protection Unit was informed by Anglian Water of an exceedence in the level of Cryptosporidium oocysts found in water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, during 19 to 24 June 2008 [25]. The reservoir supplied a population of more than 250,000 in the Northampton area. A notice advising people in the affected areas to boil all drinking water was issued on 25 June 2008 and public health messages were circulated to local health services and to the general public via the media. Those members of the public who were concerned about health risks associated with the incident were asked to ring NHS Direct for clinical advice [26]. The HPA wrote to local GPs and hospitals asking them to monitor potential patients for signs and symptoms of Cryptosporidium infection and to submit faecal specimens to the local hospital diagnostic laboratory if patients presented with diarrhoea. Samples from 34 patients where Cryptosporidium infection was identified were sent to the UK Cryptosporidium reference unit for typing. On 30 June 2008, the Cryptosporidium oocysts found in the reservoir water were confirmed as being of the rabbit genotype Cryptosporidium cuniculus [27]. Subsequently, a dead rabbit was found in a treated water tank at the water treatment works. The genotype of Cryptosporidium oocysts in the rabbit’s large bowel was indistinguishable from that of the oocysts found in the water [27]. After remediation of the water supply and distribution, the ‘boil water notice’ was lifted on 4 July and the following day the first case of cryptosporidiosis linked to the incident was identified by the reference laboratory (this case was infected with C. cuniculus). During the course of the outbreak (24 June – 18 July 2008, the dates of symptom onset in the first and last case, respectively), 23 cases of cryptosporidiosis were confirmed as being infected with C. cuniculus; one of the 23 was a secondary case. The HPA Real-time Syndromic Surveillance Team provided data in order to aid the response to this incident and the first syndromic surveillance report was circulated to the incident management team and other relevant people in the HPA on 27 June 2008. Data from the HPA/NHS Direct and HPA/QSurveillance systems were provided in a series of regular reports, initially daily and eventually weekly, until the final report on 21 August 2008. Each report included a summary interpretation and more detailed data on diarrhoea, gastroenteritis and vomiting indicators. Methods Surveillance systems HPA/NHS Direct surveillance system NHS Direct is a 24-hour nurse-led telephone helpline that provides health information and advice to the general public. Nurses use a computerised clinical decision support system – the NHS Clinical Assessment System (NHS CAS) – to handle calls. This assessment system uses approximately 200 computerised symptom-based clinical algorithms. Nurses assign the call to the most appropriate algorithm and the patient’s symptoms determine the questions asked and the action to be taken following the call (call outcome), which could be guidance on self-care or they could be referred to their GP or advised to attend a hospital emergency department. No attempt is made to provide a formal diagnosis. Daily NHS Direct data are received by the Real-time Syndromic Surveillance Team, where the number and type of calls received during the previous day are analysed and interpreted. Call proportions are calculated by age group and algorithm against the total number of calls received. HPA/QSurveillance system The HPA/QSurveillance national surveillance system was set up by the University of Nottingham, United Kingdom, and Egton Medical Information Systems (EMIS), a supplier of general practice computer systems, in collaboration with the HPA. It comprises a network of more than 3,500 general practices throughout the UK, covering more than 22 million patients (about 38% of the population [28]). Aggregated data on GP consultations for a range of indicators are automatically uploaded daily from GP practice systems to a central database. Data are routinely reported on a weekly basis; however, daily reporting is possible for specific incidents. Reports are provided at national or regional level (strategic health authority, SHA) and by local health district (primary care trust, PCT). Analysis of surveillance data NHS Direct call proportions for gastrointestinal syndromes (diarrhoea and vomiting) for the East Midlands region in England, where Northampton is situated, were examined during the outbreak (24 June – 18 July 2008) and compared with those for England and Wales. A series of control charts for diarrhoea calls are routinely used to monitor significant rises in the numbers of calls received. Control charts are calculated by assuming that calls follow a Poisson distribution with the total number of calls as an offset: a model is fitted to each region and symptom separately [29]. The model takes into account call variation caused by weekends, public holidays and the time of year – variables that can affect the number of calls received by NHS Direct. A value above the upper limit of the 99.5% confidence interval would be considered to be unusual. The seven-day moving average for diarrhoea calls was also monitored. The number and percentage of calls for diarrhoea in the East Midlands region were presented by call outcome and the number of calls in the Northampton (NN) postcode districts and in particular the number of calls in the NN11 and NN12 postcode districts, which were most affected by the incident. QSurveillance national consultation rates per 100,000 population for diarrhoea (in the age groups under five years, five years and over, and all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with rates for the same period in 2007 (data not presented). Consultation rates by region for 2008 for diarrhoea (all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with those for the East Midlands region. The gastroenteritis indicator includes all cases of diarrhoea and/or vomiting. Consultation rates and standardised incidence ratios (SIRs) – calculated using the UK as the standard population – for diarrhoea, gastroenteritis and vomiting were compared for the UK, Yorkshire and Humberside, East Midlands, Leicestershire, Northamptonshire and Rutland SHA, and Daventry and South Northants PCT, Northamptonshire Heartlands PCT and Northampton PCT. Yorkshire and Humberside was not an affected region but was included as a control. The area supplied by the Pitsford Reservoir included the three PCTs, which were all within the Leicestershire, Northamptonshire and Rutland SHA. The consultation rates and SIRs were compared for the period from week 16 to week 35 of 2008 in order to compare the rates before and after the Cryptosporidium exceedance, which took place in week 26. Estimates of excess numbers of cases of diarrhoea occurring during and following the Cryptosporidium outbreak were made by calculating the mean consultation rate over a period of five weeks before and after the incident (weeks 20–24 and weeks 31–35, respectively). For each of the three PCTs, the calculated mean rate was applied to the PCT population to estimate the number of cases that would be expected each week. The actual consultation rates for diarrhoea for weeks 25 to 30 were used to estimate the number of cases for the PCT population each week. The expected number of cases was subtracted from the estimated number of cases in the PCT population to give the estimated number of excess cases. Results HPA/NHS Direct surveillance system A peak in the number of calls for diarrhoea in the East Midlands was recorded in 25–26 June 2008, the period that coincided with the contamination incident and the associated media coverage (Figure 1). The neighbouring areas of the West Midlands, Yorkshire and the Humber, and East of England showed no increase in the number of calls for diarrhoea. The peak produced a control chart exceedance for calls for diarrhoea on 25 June 2008 (Figure 2), when the proportion of calls exceeded the upper limit of the 99.5% confidence interval. There were further confidence interval exceedances on 26 and 28 June (which were not control chart exceedances). There was no peak in calls for vomiting or control chart exceedance for these calls in the East Midlands. HPA/QSurveillance national surveillance system The East Midlands region had significantly high consultation rates for diarrhoea and gastroenteritis in week 25 (16–22 June), week 26 (23–29 June 2008, when the contamination incident was reported) and in the following four weeks. Within this region. Leicestershire, Northamptonshire and Rutland SHA had slightly raised consultation rates and significant SIRs across weeks 25 to 30 that were not seen in the neighbouring Trent SHA. At PCT level, all three of the PCTs in the area affected by the incident showed increased consultation rates for diarrhoea (Table 1) and gastroenteritis (Table 2) with SIRs significantly above the UK rate in week 26. Daventry and South Northants PCT also had a raised SIR for both indicators in week 25, and although Northamptonshire Heartlands and Northampton PCTs did not have SIRs significantly above that of the UK in week 25, the rise in consultation rates for diarrhoea and gastroenteritis began during week 25. In Northampton PCT, consultations for both diarrhoea and gastroenteritis peaked in the week following the contamination incident, week 27, returning to normal levels by week 30 (Figure 3A and 3B). A similar effect can be seen in Northampton Heartlands PCT. Daventry and South Northants PCT also showed an increase, but appeared to have consistently higher rates for both indicators. This was the area with the smallest population so the rates were more variable than in the other PCTs and we therefore interpreted these results with caution. The consultation rates for vomiting during weeks 25 to 30 in the East Midlands were not unusual at SHA or PCT level (data not presented). Discussion We have demonstrated the sensitivity of syndromic surveillance in detecting this small Cryptosporidium outbreak and the value of the surveillance in being able to describe the extent of its spread. Both the HPA/NHS Direct and HPA/QSurveillance systems showed demonstrable increases in calls and consultations for diarrhoea that were linked to the outbreak. QSurveillance consultations appeared to increase across the PCTs immediately affected but not in the surrounding area. Both the HPA/NHS Direct and HPA/QSurveillance systems showed a clear signal at the time of the incident and we were able to describe the extent of the impact on pre-primary care and primary care services. The HPA/QSurveillance system showed a rise in consultation rates for gastrointestinal symptoms that began the week before the outbreak, consistent with the period when Cryptosporidium was present in the water leaving the Pitsford Reservoir (19–24 June 2008) and with the onset of symptoms in the first outbreak case on 24 June. Although only 33 cases were identified by the outbreak investigation team, of which 23 were confirmed as having the outbreak Cryptosporidium strain, our syndromic surveillance data detected this limited outbreak. Data also suggested a more widespread increase in general gastrointestinal symptoms around the time of the outbreak, with an estimated 422 excess diarrhoea cases; these excess cases represented an increase of about 25% above normally expected levels. It is highly probable that a proportion of these excess cases may have resulted from the increased publicity surrounding the incident – for example, it is likely that media reports contributed to the large peak in calls detected by the HPA/NHS Direct surveillance system on the day the boil water notice was issued, and could also have impacted on the GP consultation rate. It has been previously shown that reporting of mumps cases is sensitive to media coverage, with a rise in clinically reported cases following newspaper reports [30]. A similar mechanism could account for some of the excess GP consultations as cases experiencing gastrointestinal symptoms may have been more likely to consult their GP, whereas in normal circumstances they would have cared for themselves at home. It is also possible that the surveillance shows outbreak-associated cases that did not come to the attention of the outbreak team, perhaps because symptoms were not sufficiently severe to warrant further investigation, or stool samples were not provided for testing. It is interesting to note that there was no demonstrable impact on the number of calls for vomiting (which is not a prominent clinical feature of cryptosporidiosis). Other common community-based pathogens such as norovirus and rotavirus were at low levels, as is normal for that time of year [31]. In this instance, public health authorities had already been alerted to a potential problem by the water company, although the extent of the outbreak was detected by syndromic surveillance. In 2003 the syndromic surveillance systems in the city of New York, United States, were able to detect an increase in diarrhoeal illness following a power outage when there was no other indication of citywide illness [32]. The New York system covers a population of nine million, but does not regularly detect localised outbreaks [33]. It has been shown previously that the HPA/NHS Direct surveillance system would be unlikely to detect a Cryptosporidium outbreak unless call volumes are high (72% chance of detection if nine-tenths of cases called NHS Direct) [29], although the value of syndromic surveillance for such outbreaks has been recognised [34]. The system detected the East Midlands Cryptosporidium outbreak that affected a smaller population than that covered by the New York system. The three PCTs affected have a combined population of around 600,000, of which just over half use GP practices reporting to QSurveillance, yet this syndromic surveillance system was able to describe an increase in consultation rates for diarrhoea and gastroenteritis around the time of the outbreak. Limitations of the data There was extensive media reporting of the incident that may have affected both the HPA/NHS Direct and HPA/QSurveillance systems and contributed to the increase in reported gastrointestinal symptoms around the time of the contamination incident. However, the rise in consultation rates for diarrhoea began before the outbreak had been detected and therefore cannot be attributed to media coverage. The HPA/NHS Direct and HPA/QSurveillance systems monitor general symptoms and so could only monitor the relevant symptoms of diarrhoea and vomiting. They are not able to detect Cryptosporidium cases, as this would require laboratory confirmation of diagnosis, so some of the estimated excess cases could be unconnected with this incident. This outbreak was discovered by other means but both the HPA/NHS Direct and HPA/ QSurveillance systems were able to describe the extent of the disease in the general population and provide reassurance that there was no widespread impact. Compared with other populations, older people and ethnic minorities are less likely to call NHS Direct [29], and although this should not prevent detection of gastrointestinal symptoms as a result of drinking water contamination as this would affect the whole population, this may reduce the signal from the system [35]. With such large surveillance systems, there will be ‘background noise’ in the data, so procedures must be in place to correctly interpret the data and set appropriate thresholds for action. Conclusion To our knowledge, this is the first time that PCT-level data from a syndromic surveillance system, the HPA/ QSurveillance national surveillance system, have been able to show the extent of such a limited outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide PCT-level data and this Cryptosporidium contamination incident demonstrates the potential usefulness of this system.

What was the event?

1495

Value of syndromic surveillance in monitoring a focal waterborne outbreak due to an unusual Cryptosporidium genotype in Northamptonshire

The United Kingdom (UK) has several national syndromic surveillance systems. The Health Protection Agency (HPA)/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data from a national telephone helpline, while the HPA/ QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems. Data from both of these systems were used to monitor a local outbreak of cryptosporidiosis that occurred following Cryptosporidium oocyst contamination of drinking water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, in June 2008. There was a peak in the number of calls to NHS Direct concerning diarrhoea that coincided with the incident. QSurveillance data for the local areas affected by the outbreak showed a significant increase in GP consultations for diarrhoea and gastroenteritis in the week of the incident but there was no increase in consultations for vomiting. A total of 33 clinical cases of cryptosporidiosis were identified in the outbreak investigation, of which 23 were confirmed as infected with the outbreak strain. However, QSurveillance data suggest that there were an estimated 422 excess diarrhoea cases during the outbreak, an increase of about 25% over baseline weekly levels. To our knowledge, this is the first time that data from a syndromic surveillance system, the HPA/QSurveillance national surveillance system, have been able to show the extent of such a small outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide data at local health district (primary care trust) level. The Cryptosporidium contamination incident described demonstrates the potential usefulness of this information, as it is unusual for syndromic surveillance systems to be able to help monitor such a small-scale outbreak. Introduction As syndromic surveillance systems usually capture data already collected for other purposes, and monitor generic symptoms and/or clinically diagnosed disease, they provide information at an earlier stage of illness (compared with laboratory-confirmed diagnoses), so that action can be taken in time to substantially reduce the impact of disease. Some systems, for example, the Royal College of General Practitioners Weekly Returns Service, are now well established, with many years of historical data that allow monitoring of longer-term disease trends [1]. They have the ability to provide early warning of, for example, seasonal rises in influenza and can trigger public health action, such as a recommendation to prescribe antiviral drugs in line with national guidance [2-4]. They can also provide reassurance to incident response teams and the general public that an incident has not caused adverse health effects – for example, following an explosion at the Buncefield oil storage depot in Hemel Hempstead, United Kingdom (UK), in 2005, syndromic surveillance confirmed that there were no unusual rises in community-based morbidity linked to the incident [5]; following the eruption of the Eyjafjallajökull volcano in Iceland in April 2010 similar assurance was given about lack of impact on community morbidity [6]. Health departments are increasingly expected to monitor health effects of natural events such as heat wave or flooding, or implement surveillance – of which syndromic surveillance plays a major role – for mass gatherings such as the Olympics or football World Cup [7-9]. Systems in France, Australia and Taiwan use data from emergency departments [10-12], a Canadian system uses over-the-counter pharmacy sales [13,14], and in the Netherlands data from both syndromic and surrogate data sources, such as employee absence records and prescription medications dispensed by pharmacies, are included in surveillance systems [15,16]. Currently systems based on Internet searches via search engines or on queries submitted to medical websites are being developed [17,18]. In the UK, the HPA/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data collected from the NHS Direct telephone helpline [19], while the HPA/QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems [20]. The HPA Real-time Syndromic Surveillance Team is a small team that coordinates a number of syndromic surveillance systems within the HPA and takes a lead for syndromic surveillance in England [21]. This paper describes the support provided by the team to the local incident management team during a local cryptosporidiosis outbreak and shows the use of syndromic surveillance in monitoring the extent of an outbreak using the HPA/NHS Direct and HPA/QSurveillance national surveillance systems. Cryptosporidiosis Cryptosporidium is a protozoan parasite that can cause an infection in people, cattle and sometimes other animals [22]. Cryptosporidiosis is most common in children aged between one and five years, but it can affect all ages. Those with impaired immune systems are likely to be most seriously affected. Symptoms usually appear between three and 12 days after initial exposure and include watery diarrhoea, stomach pains, dehydration and fever. In its transmissible form, called an oocyst, the parasite is protected by an outer shell, which allows it to survive in the environment for a long time. Transmission occurs most often via the faeco-oral route through person-to-person or animalto-person contact, but people may also be infected by consuming contaminated water or food or by swimming in contaminated water. Although uncommon, the largest outbreaks have occurred following contamination of drinking water [23,24]. Normal chlorine disinfection procedures do not kill the oocysts, so they are removed by filtration and water companies carry out routine monitoring of treated water. Description of the incident On 25 June 2008 the local Health Protection Unit was informed by Anglian Water of an exceedence in the level of Cryptosporidium oocysts found in water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, during 19 to 24 June 2008 [25]. The reservoir supplied a population of more than 250,000 in the Northampton area. A notice advising people in the affected areas to boil all drinking water was issued on 25 June 2008 and public health messages were circulated to local health services and to the general public via the media. Those members of the public who were concerned about health risks associated with the incident were asked to ring NHS Direct for clinical advice [26]. The HPA wrote to local GPs and hospitals asking them to monitor potential patients for signs and symptoms of Cryptosporidium infection and to submit faecal specimens to the local hospital diagnostic laboratory if patients presented with diarrhoea. Samples from 34 patients where Cryptosporidium infection was identified were sent to the UK Cryptosporidium reference unit for typing. On 30 June 2008, the Cryptosporidium oocysts found in the reservoir water were confirmed as being of the rabbit genotype Cryptosporidium cuniculus [27]. Subsequently, a dead rabbit was found in a treated water tank at the water treatment works. The genotype of Cryptosporidium oocysts in the rabbit’s large bowel was indistinguishable from that of the oocysts found in the water [27]. After remediation of the water supply and distribution, the ‘boil water notice’ was lifted on 4 July and the following day the first case of cryptosporidiosis linked to the incident was identified by the reference laboratory (this case was infected with C. cuniculus). During the course of the outbreak (24 June – 18 July 2008, the dates of symptom onset in the first and last case, respectively), 23 cases of cryptosporidiosis were confirmed as being infected with C. cuniculus; one of the 23 was a secondary case. The HPA Real-time Syndromic Surveillance Team provided data in order to aid the response to this incident and the first syndromic surveillance report was circulated to the incident management team and other relevant people in the HPA on 27 June 2008. Data from the HPA/NHS Direct and HPA/QSurveillance systems were provided in a series of regular reports, initially daily and eventually weekly, until the final report on 21 August 2008. Each report included a summary interpretation and more detailed data on diarrhoea, gastroenteritis and vomiting indicators. Methods Surveillance systems HPA/NHS Direct surveillance system NHS Direct is a 24-hour nurse-led telephone helpline that provides health information and advice to the general public. Nurses use a computerised clinical decision support system – the NHS Clinical Assessment System (NHS CAS) – to handle calls. This assessment system uses approximately 200 computerised symptom-based clinical algorithms. Nurses assign the call to the most appropriate algorithm and the patient’s symptoms determine the questions asked and the action to be taken following the call (call outcome), which could be guidance on self-care or they could be referred to their GP or advised to attend a hospital emergency department. No attempt is made to provide a formal diagnosis. Daily NHS Direct data are received by the Real-time Syndromic Surveillance Team, where the number and type of calls received during the previous day are analysed and interpreted. Call proportions are calculated by age group and algorithm against the total number of calls received. HPA/QSurveillance system The HPA/QSurveillance national surveillance system was set up by the University of Nottingham, United Kingdom, and Egton Medical Information Systems (EMIS), a supplier of general practice computer systems, in collaboration with the HPA. It comprises a network of more than 3,500 general practices throughout the UK, covering more than 22 million patients (about 38% of the population [28]). Aggregated data on GP consultations for a range of indicators are automatically uploaded daily from GP practice systems to a central database. Data are routinely reported on a weekly basis; however, daily reporting is possible for specific incidents. Reports are provided at national or regional level (strategic health authority, SHA) and by local health district (primary care trust, PCT). Analysis of surveillance data NHS Direct call proportions for gastrointestinal syndromes (diarrhoea and vomiting) for the East Midlands region in England, where Northampton is situated, were examined during the outbreak (24 June – 18 July 2008) and compared with those for England and Wales. A series of control charts for diarrhoea calls are routinely used to monitor significant rises in the numbers of calls received. Control charts are calculated by assuming that calls follow a Poisson distribution with the total number of calls as an offset: a model is fitted to each region and symptom separately [29]. The model takes into account call variation caused by weekends, public holidays and the time of year – variables that can affect the number of calls received by NHS Direct. A value above the upper limit of the 99.5% confidence interval would be considered to be unusual. The seven-day moving average for diarrhoea calls was also monitored. The number and percentage of calls for diarrhoea in the East Midlands region were presented by call outcome and the number of calls in the Northampton (NN) postcode districts and in particular the number of calls in the NN11 and NN12 postcode districts, which were most affected by the incident. QSurveillance national consultation rates per 100,000 population for diarrhoea (in the age groups under five years, five years and over, and all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with rates for the same period in 2007 (data not presented). Consultation rates by region for 2008 for diarrhoea (all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with those for the East Midlands region. The gastroenteritis indicator includes all cases of diarrhoea and/or vomiting. Consultation rates and standardised incidence ratios (SIRs) – calculated using the UK as the standard population – for diarrhoea, gastroenteritis and vomiting were compared for the UK, Yorkshire and Humberside, East Midlands, Leicestershire, Northamptonshire and Rutland SHA, and Daventry and South Northants PCT, Northamptonshire Heartlands PCT and Northampton PCT. Yorkshire and Humberside was not an affected region but was included as a control. The area supplied by the Pitsford Reservoir included the three PCTs, which were all within the Leicestershire, Northamptonshire and Rutland SHA. The consultation rates and SIRs were compared for the period from week 16 to week 35 of 2008 in order to compare the rates before and after the Cryptosporidium exceedance, which took place in week 26. Estimates of excess numbers of cases of diarrhoea occurring during and following the Cryptosporidium outbreak were made by calculating the mean consultation rate over a period of five weeks before and after the incident (weeks 20–24 and weeks 31–35, respectively). For each of the three PCTs, the calculated mean rate was applied to the PCT population to estimate the number of cases that would be expected each week. The actual consultation rates for diarrhoea for weeks 25 to 30 were used to estimate the number of cases for the PCT population each week. The expected number of cases was subtracted from the estimated number of cases in the PCT population to give the estimated number of excess cases. Results HPA/NHS Direct surveillance system A peak in the number of calls for diarrhoea in the East Midlands was recorded in 25–26 June 2008, the period that coincided with the contamination incident and the associated media coverage (Figure 1). The neighbouring areas of the West Midlands, Yorkshire and the Humber, and East of England showed no increase in the number of calls for diarrhoea. The peak produced a control chart exceedance for calls for diarrhoea on 25 June 2008 (Figure 2), when the proportion of calls exceeded the upper limit of the 99.5% confidence interval. There were further confidence interval exceedances on 26 and 28 June (which were not control chart exceedances). There was no peak in calls for vomiting or control chart exceedance for these calls in the East Midlands. HPA/QSurveillance national surveillance system The East Midlands region had significantly high consultation rates for diarrhoea and gastroenteritis in week 25 (16–22 June), week 26 (23–29 June 2008, when the contamination incident was reported) and in the following four weeks. Within this region. Leicestershire, Northamptonshire and Rutland SHA had slightly raised consultation rates and significant SIRs across weeks 25 to 30 that were not seen in the neighbouring Trent SHA. At PCT level, all three of the PCTs in the area affected by the incident showed increased consultation rates for diarrhoea (Table 1) and gastroenteritis (Table 2) with SIRs significantly above the UK rate in week 26. Daventry and South Northants PCT also had a raised SIR for both indicators in week 25, and although Northamptonshire Heartlands and Northampton PCTs did not have SIRs significantly above that of the UK in week 25, the rise in consultation rates for diarrhoea and gastroenteritis began during week 25. In Northampton PCT, consultations for both diarrhoea and gastroenteritis peaked in the week following the contamination incident, week 27, returning to normal levels by week 30 (Figure 3A and 3B). A similar effect can be seen in Northampton Heartlands PCT. Daventry and South Northants PCT also showed an increase, but appeared to have consistently higher rates for both indicators. This was the area with the smallest population so the rates were more variable than in the other PCTs and we therefore interpreted these results with caution. The consultation rates for vomiting during weeks 25 to 30 in the East Midlands were not unusual at SHA or PCT level (data not presented). Discussion We have demonstrated the sensitivity of syndromic surveillance in detecting this small Cryptosporidium outbreak and the value of the surveillance in being able to describe the extent of its spread. Both the HPA/NHS Direct and HPA/QSurveillance systems showed demonstrable increases in calls and consultations for diarrhoea that were linked to the outbreak. QSurveillance consultations appeared to increase across the PCTs immediately affected but not in the surrounding area. Both the HPA/NHS Direct and HPA/QSurveillance systems showed a clear signal at the time of the incident and we were able to describe the extent of the impact on pre-primary care and primary care services. The HPA/QSurveillance system showed a rise in consultation rates for gastrointestinal symptoms that began the week before the outbreak, consistent with the period when Cryptosporidium was present in the water leaving the Pitsford Reservoir (19–24 June 2008) and with the onset of symptoms in the first outbreak case on 24 June. Although only 33 cases were identified by the outbreak investigation team, of which 23 were confirmed as having the outbreak Cryptosporidium strain, our syndromic surveillance data detected this limited outbreak. Data also suggested a more widespread increase in general gastrointestinal symptoms around the time of the outbreak, with an estimated 422 excess diarrhoea cases; these excess cases represented an increase of about 25% above normally expected levels. It is highly probable that a proportion of these excess cases may have resulted from the increased publicity surrounding the incident – for example, it is likely that media reports contributed to the large peak in calls detected by the HPA/NHS Direct surveillance system on the day the boil water notice was issued, and could also have impacted on the GP consultation rate. It has been previously shown that reporting of mumps cases is sensitive to media coverage, with a rise in clinically reported cases following newspaper reports [30]. A similar mechanism could account for some of the excess GP consultations as cases experiencing gastrointestinal symptoms may have been more likely to consult their GP, whereas in normal circumstances they would have cared for themselves at home. It is also possible that the surveillance shows outbreak-associated cases that did not come to the attention of the outbreak team, perhaps because symptoms were not sufficiently severe to warrant further investigation, or stool samples were not provided for testing. It is interesting to note that there was no demonstrable impact on the number of calls for vomiting (which is not a prominent clinical feature of cryptosporidiosis). Other common community-based pathogens such as norovirus and rotavirus were at low levels, as is normal for that time of year [31]. In this instance, public health authorities had already been alerted to a potential problem by the water company, although the extent of the outbreak was detected by syndromic surveillance. In 2003 the syndromic surveillance systems in the city of New York, United States, were able to detect an increase in diarrhoeal illness following a power outage when there was no other indication of citywide illness [32]. The New York system covers a population of nine million, but does not regularly detect localised outbreaks [33]. It has been shown previously that the HPA/NHS Direct surveillance system would be unlikely to detect a Cryptosporidium outbreak unless call volumes are high (72% chance of detection if nine-tenths of cases called NHS Direct) [29], although the value of syndromic surveillance for such outbreaks has been recognised [34]. The system detected the East Midlands Cryptosporidium outbreak that affected a smaller population than that covered by the New York system. The three PCTs affected have a combined population of around 600,000, of which just over half use GP practices reporting to QSurveillance, yet this syndromic surveillance system was able to describe an increase in consultation rates for diarrhoea and gastroenteritis around the time of the outbreak. Limitations of the data There was extensive media reporting of the incident that may have affected both the HPA/NHS Direct and HPA/QSurveillance systems and contributed to the increase in reported gastrointestinal symptoms around the time of the contamination incident. However, the rise in consultation rates for diarrhoea began before the outbreak had been detected and therefore cannot be attributed to media coverage. The HPA/NHS Direct and HPA/QSurveillance systems monitor general symptoms and so could only monitor the relevant symptoms of diarrhoea and vomiting. They are not able to detect Cryptosporidium cases, as this would require laboratory confirmation of diagnosis, so some of the estimated excess cases could be unconnected with this incident. This outbreak was discovered by other means but both the HPA/NHS Direct and HPA/ QSurveillance systems were able to describe the extent of the disease in the general population and provide reassurance that there was no widespread impact. Compared with other populations, older people and ethnic minorities are less likely to call NHS Direct [29], and although this should not prevent detection of gastrointestinal symptoms as a result of drinking water contamination as this would affect the whole population, this may reduce the signal from the system [35]. With such large surveillance systems, there will be ‘background noise’ in the data, so procedures must be in place to correctly interpret the data and set appropriate thresholds for action. Conclusion To our knowledge, this is the first time that PCT-level data from a syndromic surveillance system, the HPA/ QSurveillance national surveillance system, have been able to show the extent of such a limited outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide PCT-level data and this Cryptosporidium contamination incident demonstrates the potential usefulness of this system.

When did this happen?

1496

Value of syndromic surveillance in monitoring a focal waterborne outbreak due to an unusual Cryptosporidium genotype in Northamptonshire

The United Kingdom (UK) has several national syndromic surveillance systems. The Health Protection Agency (HPA)/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data from a national telephone helpline, while the HPA/ QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems. Data from both of these systems were used to monitor a local outbreak of cryptosporidiosis that occurred following Cryptosporidium oocyst contamination of drinking water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, in June 2008. There was a peak in the number of calls to NHS Direct concerning diarrhoea that coincided with the incident. QSurveillance data for the local areas affected by the outbreak showed a significant increase in GP consultations for diarrhoea and gastroenteritis in the week of the incident but there was no increase in consultations for vomiting. A total of 33 clinical cases of cryptosporidiosis were identified in the outbreak investigation, of which 23 were confirmed as infected with the outbreak strain. However, QSurveillance data suggest that there were an estimated 422 excess diarrhoea cases during the outbreak, an increase of about 25% over baseline weekly levels. To our knowledge, this is the first time that data from a syndromic surveillance system, the HPA/QSurveillance national surveillance system, have been able to show the extent of such a small outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide data at local health district (primary care trust) level. The Cryptosporidium contamination incident described demonstrates the potential usefulness of this information, as it is unusual for syndromic surveillance systems to be able to help monitor such a small-scale outbreak. Introduction As syndromic surveillance systems usually capture data already collected for other purposes, and monitor generic symptoms and/or clinically diagnosed disease, they provide information at an earlier stage of illness (compared with laboratory-confirmed diagnoses), so that action can be taken in time to substantially reduce the impact of disease. Some systems, for example, the Royal College of General Practitioners Weekly Returns Service, are now well established, with many years of historical data that allow monitoring of longer-term disease trends [1]. They have the ability to provide early warning of, for example, seasonal rises in influenza and can trigger public health action, such as a recommendation to prescribe antiviral drugs in line with national guidance [2-4]. They can also provide reassurance to incident response teams and the general public that an incident has not caused adverse health effects – for example, following an explosion at the Buncefield oil storage depot in Hemel Hempstead, United Kingdom (UK), in 2005, syndromic surveillance confirmed that there were no unusual rises in community-based morbidity linked to the incident [5]; following the eruption of the Eyjafjallajökull volcano in Iceland in April 2010 similar assurance was given about lack of impact on community morbidity [6]. Health departments are increasingly expected to monitor health effects of natural events such as heat wave or flooding, or implement surveillance – of which syndromic surveillance plays a major role – for mass gatherings such as the Olympics or football World Cup [7-9]. Systems in France, Australia and Taiwan use data from emergency departments [10-12], a Canadian system uses over-the-counter pharmacy sales [13,14], and in the Netherlands data from both syndromic and surrogate data sources, such as employee absence records and prescription medications dispensed by pharmacies, are included in surveillance systems [15,16]. Currently systems based on Internet searches via search engines or on queries submitted to medical websites are being developed [17,18]. In the UK, the HPA/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data collected from the NHS Direct telephone helpline [19], while the HPA/QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems [20]. The HPA Real-time Syndromic Surveillance Team is a small team that coordinates a number of syndromic surveillance systems within the HPA and takes a lead for syndromic surveillance in England [21]. This paper describes the support provided by the team to the local incident management team during a local cryptosporidiosis outbreak and shows the use of syndromic surveillance in monitoring the extent of an outbreak using the HPA/NHS Direct and HPA/QSurveillance national surveillance systems. Cryptosporidiosis Cryptosporidium is a protozoan parasite that can cause an infection in people, cattle and sometimes other animals [22]. Cryptosporidiosis is most common in children aged between one and five years, but it can affect all ages. Those with impaired immune systems are likely to be most seriously affected. Symptoms usually appear between three and 12 days after initial exposure and include watery diarrhoea, stomach pains, dehydration and fever. In its transmissible form, called an oocyst, the parasite is protected by an outer shell, which allows it to survive in the environment for a long time. Transmission occurs most often via the faeco-oral route through person-to-person or animalto-person contact, but people may also be infected by consuming contaminated water or food or by swimming in contaminated water. Although uncommon, the largest outbreaks have occurred following contamination of drinking water [23,24]. Normal chlorine disinfection procedures do not kill the oocysts, so they are removed by filtration and water companies carry out routine monitoring of treated water. Description of the incident On 25 June 2008 the local Health Protection Unit was informed by Anglian Water of an exceedence in the level of Cryptosporidium oocysts found in water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, during 19 to 24 June 2008 [25]. The reservoir supplied a population of more than 250,000 in the Northampton area. A notice advising people in the affected areas to boil all drinking water was issued on 25 June 2008 and public health messages were circulated to local health services and to the general public via the media. Those members of the public who were concerned about health risks associated with the incident were asked to ring NHS Direct for clinical advice [26]. The HPA wrote to local GPs and hospitals asking them to monitor potential patients for signs and symptoms of Cryptosporidium infection and to submit faecal specimens to the local hospital diagnostic laboratory if patients presented with diarrhoea. Samples from 34 patients where Cryptosporidium infection was identified were sent to the UK Cryptosporidium reference unit for typing. On 30 June 2008, the Cryptosporidium oocysts found in the reservoir water were confirmed as being of the rabbit genotype Cryptosporidium cuniculus [27]. Subsequently, a dead rabbit was found in a treated water tank at the water treatment works. The genotype of Cryptosporidium oocysts in the rabbit’s large bowel was indistinguishable from that of the oocysts found in the water [27]. After remediation of the water supply and distribution, the ‘boil water notice’ was lifted on 4 July and the following day the first case of cryptosporidiosis linked to the incident was identified by the reference laboratory (this case was infected with C. cuniculus). During the course of the outbreak (24 June – 18 July 2008, the dates of symptom onset in the first and last case, respectively), 23 cases of cryptosporidiosis were confirmed as being infected with C. cuniculus; one of the 23 was a secondary case. The HPA Real-time Syndromic Surveillance Team provided data in order to aid the response to this incident and the first syndromic surveillance report was circulated to the incident management team and other relevant people in the HPA on 27 June 2008. Data from the HPA/NHS Direct and HPA/QSurveillance systems were provided in a series of regular reports, initially daily and eventually weekly, until the final report on 21 August 2008. Each report included a summary interpretation and more detailed data on diarrhoea, gastroenteritis and vomiting indicators. Methods Surveillance systems HPA/NHS Direct surveillance system NHS Direct is a 24-hour nurse-led telephone helpline that provides health information and advice to the general public. Nurses use a computerised clinical decision support system – the NHS Clinical Assessment System (NHS CAS) – to handle calls. This assessment system uses approximately 200 computerised symptom-based clinical algorithms. Nurses assign the call to the most appropriate algorithm and the patient’s symptoms determine the questions asked and the action to be taken following the call (call outcome), which could be guidance on self-care or they could be referred to their GP or advised to attend a hospital emergency department. No attempt is made to provide a formal diagnosis. Daily NHS Direct data are received by the Real-time Syndromic Surveillance Team, where the number and type of calls received during the previous day are analysed and interpreted. Call proportions are calculated by age group and algorithm against the total number of calls received. HPA/QSurveillance system The HPA/QSurveillance national surveillance system was set up by the University of Nottingham, United Kingdom, and Egton Medical Information Systems (EMIS), a supplier of general practice computer systems, in collaboration with the HPA. It comprises a network of more than 3,500 general practices throughout the UK, covering more than 22 million patients (about 38% of the population [28]). Aggregated data on GP consultations for a range of indicators are automatically uploaded daily from GP practice systems to a central database. Data are routinely reported on a weekly basis; however, daily reporting is possible for specific incidents. Reports are provided at national or regional level (strategic health authority, SHA) and by local health district (primary care trust, PCT). Analysis of surveillance data NHS Direct call proportions for gastrointestinal syndromes (diarrhoea and vomiting) for the East Midlands region in England, where Northampton is situated, were examined during the outbreak (24 June – 18 July 2008) and compared with those for England and Wales. A series of control charts for diarrhoea calls are routinely used to monitor significant rises in the numbers of calls received. Control charts are calculated by assuming that calls follow a Poisson distribution with the total number of calls as an offset: a model is fitted to each region and symptom separately [29]. The model takes into account call variation caused by weekends, public holidays and the time of year – variables that can affect the number of calls received by NHS Direct. A value above the upper limit of the 99.5% confidence interval would be considered to be unusual. The seven-day moving average for diarrhoea calls was also monitored. The number and percentage of calls for diarrhoea in the East Midlands region were presented by call outcome and the number of calls in the Northampton (NN) postcode districts and in particular the number of calls in the NN11 and NN12 postcode districts, which were most affected by the incident. QSurveillance national consultation rates per 100,000 population for diarrhoea (in the age groups under five years, five years and over, and all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with rates for the same period in 2007 (data not presented). Consultation rates by region for 2008 for diarrhoea (all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with those for the East Midlands region. The gastroenteritis indicator includes all cases of diarrhoea and/or vomiting. Consultation rates and standardised incidence ratios (SIRs) – calculated using the UK as the standard population – for diarrhoea, gastroenteritis and vomiting were compared for the UK, Yorkshire and Humberside, East Midlands, Leicestershire, Northamptonshire and Rutland SHA, and Daventry and South Northants PCT, Northamptonshire Heartlands PCT and Northampton PCT. Yorkshire and Humberside was not an affected region but was included as a control. The area supplied by the Pitsford Reservoir included the three PCTs, which were all within the Leicestershire, Northamptonshire and Rutland SHA. The consultation rates and SIRs were compared for the period from week 16 to week 35 of 2008 in order to compare the rates before and after the Cryptosporidium exceedance, which took place in week 26. Estimates of excess numbers of cases of diarrhoea occurring during and following the Cryptosporidium outbreak were made by calculating the mean consultation rate over a period of five weeks before and after the incident (weeks 20–24 and weeks 31–35, respectively). For each of the three PCTs, the calculated mean rate was applied to the PCT population to estimate the number of cases that would be expected each week. The actual consultation rates for diarrhoea for weeks 25 to 30 were used to estimate the number of cases for the PCT population each week. The expected number of cases was subtracted from the estimated number of cases in the PCT population to give the estimated number of excess cases. Results HPA/NHS Direct surveillance system A peak in the number of calls for diarrhoea in the East Midlands was recorded in 25–26 June 2008, the period that coincided with the contamination incident and the associated media coverage (Figure 1). The neighbouring areas of the West Midlands, Yorkshire and the Humber, and East of England showed no increase in the number of calls for diarrhoea. The peak produced a control chart exceedance for calls for diarrhoea on 25 June 2008 (Figure 2), when the proportion of calls exceeded the upper limit of the 99.5% confidence interval. There were further confidence interval exceedances on 26 and 28 June (which were not control chart exceedances). There was no peak in calls for vomiting or control chart exceedance for these calls in the East Midlands. HPA/QSurveillance national surveillance system The East Midlands region had significantly high consultation rates for diarrhoea and gastroenteritis in week 25 (16–22 June), week 26 (23–29 June 2008, when the contamination incident was reported) and in the following four weeks. Within this region. Leicestershire, Northamptonshire and Rutland SHA had slightly raised consultation rates and significant SIRs across weeks 25 to 30 that were not seen in the neighbouring Trent SHA. At PCT level, all three of the PCTs in the area affected by the incident showed increased consultation rates for diarrhoea (Table 1) and gastroenteritis (Table 2) with SIRs significantly above the UK rate in week 26. Daventry and South Northants PCT also had a raised SIR for both indicators in week 25, and although Northamptonshire Heartlands and Northampton PCTs did not have SIRs significantly above that of the UK in week 25, the rise in consultation rates for diarrhoea and gastroenteritis began during week 25. In Northampton PCT, consultations for both diarrhoea and gastroenteritis peaked in the week following the contamination incident, week 27, returning to normal levels by week 30 (Figure 3A and 3B). A similar effect can be seen in Northampton Heartlands PCT. Daventry and South Northants PCT also showed an increase, but appeared to have consistently higher rates for both indicators. This was the area with the smallest population so the rates were more variable than in the other PCTs and we therefore interpreted these results with caution. The consultation rates for vomiting during weeks 25 to 30 in the East Midlands were not unusual at SHA or PCT level (data not presented). Discussion We have demonstrated the sensitivity of syndromic surveillance in detecting this small Cryptosporidium outbreak and the value of the surveillance in being able to describe the extent of its spread. Both the HPA/NHS Direct and HPA/QSurveillance systems showed demonstrable increases in calls and consultations for diarrhoea that were linked to the outbreak. QSurveillance consultations appeared to increase across the PCTs immediately affected but not in the surrounding area. Both the HPA/NHS Direct and HPA/QSurveillance systems showed a clear signal at the time of the incident and we were able to describe the extent of the impact on pre-primary care and primary care services. The HPA/QSurveillance system showed a rise in consultation rates for gastrointestinal symptoms that began the week before the outbreak, consistent with the period when Cryptosporidium was present in the water leaving the Pitsford Reservoir (19–24 June 2008) and with the onset of symptoms in the first outbreak case on 24 June. Although only 33 cases were identified by the outbreak investigation team, of which 23 were confirmed as having the outbreak Cryptosporidium strain, our syndromic surveillance data detected this limited outbreak. Data also suggested a more widespread increase in general gastrointestinal symptoms around the time of the outbreak, with an estimated 422 excess diarrhoea cases; these excess cases represented an increase of about 25% above normally expected levels. It is highly probable that a proportion of these excess cases may have resulted from the increased publicity surrounding the incident – for example, it is likely that media reports contributed to the large peak in calls detected by the HPA/NHS Direct surveillance system on the day the boil water notice was issued, and could also have impacted on the GP consultation rate. It has been previously shown that reporting of mumps cases is sensitive to media coverage, with a rise in clinically reported cases following newspaper reports [30]. A similar mechanism could account for some of the excess GP consultations as cases experiencing gastrointestinal symptoms may have been more likely to consult their GP, whereas in normal circumstances they would have cared for themselves at home. It is also possible that the surveillance shows outbreak-associated cases that did not come to the attention of the outbreak team, perhaps because symptoms were not sufficiently severe to warrant further investigation, or stool samples were not provided for testing. It is interesting to note that there was no demonstrable impact on the number of calls for vomiting (which is not a prominent clinical feature of cryptosporidiosis). Other common community-based pathogens such as norovirus and rotavirus were at low levels, as is normal for that time of year [31]. In this instance, public health authorities had already been alerted to a potential problem by the water company, although the extent of the outbreak was detected by syndromic surveillance. In 2003 the syndromic surveillance systems in the city of New York, United States, were able to detect an increase in diarrhoeal illness following a power outage when there was no other indication of citywide illness [32]. The New York system covers a population of nine million, but does not regularly detect localised outbreaks [33]. It has been shown previously that the HPA/NHS Direct surveillance system would be unlikely to detect a Cryptosporidium outbreak unless call volumes are high (72% chance of detection if nine-tenths of cases called NHS Direct) [29], although the value of syndromic surveillance for such outbreaks has been recognised [34]. The system detected the East Midlands Cryptosporidium outbreak that affected a smaller population than that covered by the New York system. The three PCTs affected have a combined population of around 600,000, of which just over half use GP practices reporting to QSurveillance, yet this syndromic surveillance system was able to describe an increase in consultation rates for diarrhoea and gastroenteritis around the time of the outbreak. Limitations of the data There was extensive media reporting of the incident that may have affected both the HPA/NHS Direct and HPA/QSurveillance systems and contributed to the increase in reported gastrointestinal symptoms around the time of the contamination incident. However, the rise in consultation rates for diarrhoea began before the outbreak had been detected and therefore cannot be attributed to media coverage. The HPA/NHS Direct and HPA/QSurveillance systems monitor general symptoms and so could only monitor the relevant symptoms of diarrhoea and vomiting. They are not able to detect Cryptosporidium cases, as this would require laboratory confirmation of diagnosis, so some of the estimated excess cases could be unconnected with this incident. This outbreak was discovered by other means but both the HPA/NHS Direct and HPA/ QSurveillance systems were able to describe the extent of the disease in the general population and provide reassurance that there was no widespread impact. Compared with other populations, older people and ethnic minorities are less likely to call NHS Direct [29], and although this should not prevent detection of gastrointestinal symptoms as a result of drinking water contamination as this would affect the whole population, this may reduce the signal from the system [35]. With such large surveillance systems, there will be ‘background noise’ in the data, so procedures must be in place to correctly interpret the data and set appropriate thresholds for action. Conclusion To our knowledge, this is the first time that PCT-level data from a syndromic surveillance system, the HPA/ QSurveillance national surveillance system, have been able to show the extent of such a limited outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide PCT-level data and this Cryptosporidium contamination incident demonstrates the potential usefulness of this system.

When did this event start?

1497

Value of syndromic surveillance in monitoring a focal waterborne outbreak due to an unusual Cryptosporidium genotype in Northamptonshire

The United Kingdom (UK) has several national syndromic surveillance systems. The Health Protection Agency (HPA)/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data from a national telephone helpline, while the HPA/ QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems. Data from both of these systems were used to monitor a local outbreak of cryptosporidiosis that occurred following Cryptosporidium oocyst contamination of drinking water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, in June 2008. There was a peak in the number of calls to NHS Direct concerning diarrhoea that coincided with the incident. QSurveillance data for the local areas affected by the outbreak showed a significant increase in GP consultations for diarrhoea and gastroenteritis in the week of the incident but there was no increase in consultations for vomiting. A total of 33 clinical cases of cryptosporidiosis were identified in the outbreak investigation, of which 23 were confirmed as infected with the outbreak strain. However, QSurveillance data suggest that there were an estimated 422 excess diarrhoea cases during the outbreak, an increase of about 25% over baseline weekly levels. To our knowledge, this is the first time that data from a syndromic surveillance system, the HPA/QSurveillance national surveillance system, have been able to show the extent of such a small outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide data at local health district (primary care trust) level. The Cryptosporidium contamination incident described demonstrates the potential usefulness of this information, as it is unusual for syndromic surveillance systems to be able to help monitor such a small-scale outbreak. Introduction As syndromic surveillance systems usually capture data already collected for other purposes, and monitor generic symptoms and/or clinically diagnosed disease, they provide information at an earlier stage of illness (compared with laboratory-confirmed diagnoses), so that action can be taken in time to substantially reduce the impact of disease. Some systems, for example, the Royal College of General Practitioners Weekly Returns Service, are now well established, with many years of historical data that allow monitoring of longer-term disease trends [1]. They have the ability to provide early warning of, for example, seasonal rises in influenza and can trigger public health action, such as a recommendation to prescribe antiviral drugs in line with national guidance [2-4]. They can also provide reassurance to incident response teams and the general public that an incident has not caused adverse health effects – for example, following an explosion at the Buncefield oil storage depot in Hemel Hempstead, United Kingdom (UK), in 2005, syndromic surveillance confirmed that there were no unusual rises in community-based morbidity linked to the incident [5]; following the eruption of the Eyjafjallajökull volcano in Iceland in April 2010 similar assurance was given about lack of impact on community morbidity [6]. Health departments are increasingly expected to monitor health effects of natural events such as heat wave or flooding, or implement surveillance – of which syndromic surveillance plays a major role – for mass gatherings such as the Olympics or football World Cup [7-9]. Systems in France, Australia and Taiwan use data from emergency departments [10-12], a Canadian system uses over-the-counter pharmacy sales [13,14], and in the Netherlands data from both syndromic and surrogate data sources, such as employee absence records and prescription medications dispensed by pharmacies, are included in surveillance systems [15,16]. Currently systems based on Internet searches via search engines or on queries submitted to medical websites are being developed [17,18]. In the UK, the HPA/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data collected from the NHS Direct telephone helpline [19], while the HPA/QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems [20]. The HPA Real-time Syndromic Surveillance Team is a small team that coordinates a number of syndromic surveillance systems within the HPA and takes a lead for syndromic surveillance in England [21]. This paper describes the support provided by the team to the local incident management team during a local cryptosporidiosis outbreak and shows the use of syndromic surveillance in monitoring the extent of an outbreak using the HPA/NHS Direct and HPA/QSurveillance national surveillance systems. Cryptosporidiosis Cryptosporidium is a protozoan parasite that can cause an infection in people, cattle and sometimes other animals [22]. Cryptosporidiosis is most common in children aged between one and five years, but it can affect all ages. Those with impaired immune systems are likely to be most seriously affected. Symptoms usually appear between three and 12 days after initial exposure and include watery diarrhoea, stomach pains, dehydration and fever. In its transmissible form, called an oocyst, the parasite is protected by an outer shell, which allows it to survive in the environment for a long time. Transmission occurs most often via the faeco-oral route through person-to-person or animalto-person contact, but people may also be infected by consuming contaminated water or food or by swimming in contaminated water. Although uncommon, the largest outbreaks have occurred following contamination of drinking water [23,24]. Normal chlorine disinfection procedures do not kill the oocysts, so they are removed by filtration and water companies carry out routine monitoring of treated water. Description of the incident On 25 June 2008 the local Health Protection Unit was informed by Anglian Water of an exceedence in the level of Cryptosporidium oocysts found in water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, during 19 to 24 June 2008 [25]. The reservoir supplied a population of more than 250,000 in the Northampton area. A notice advising people in the affected areas to boil all drinking water was issued on 25 June 2008 and public health messages were circulated to local health services and to the general public via the media. Those members of the public who were concerned about health risks associated with the incident were asked to ring NHS Direct for clinical advice [26]. The HPA wrote to local GPs and hospitals asking them to monitor potential patients for signs and symptoms of Cryptosporidium infection and to submit faecal specimens to the local hospital diagnostic laboratory if patients presented with diarrhoea. Samples from 34 patients where Cryptosporidium infection was identified were sent to the UK Cryptosporidium reference unit for typing. On 30 June 2008, the Cryptosporidium oocysts found in the reservoir water were confirmed as being of the rabbit genotype Cryptosporidium cuniculus [27]. Subsequently, a dead rabbit was found in a treated water tank at the water treatment works. The genotype of Cryptosporidium oocysts in the rabbit’s large bowel was indistinguishable from that of the oocysts found in the water [27]. After remediation of the water supply and distribution, the ‘boil water notice’ was lifted on 4 July and the following day the first case of cryptosporidiosis linked to the incident was identified by the reference laboratory (this case was infected with C. cuniculus). During the course of the outbreak (24 June – 18 July 2008, the dates of symptom onset in the first and last case, respectively), 23 cases of cryptosporidiosis were confirmed as being infected with C. cuniculus; one of the 23 was a secondary case. The HPA Real-time Syndromic Surveillance Team provided data in order to aid the response to this incident and the first syndromic surveillance report was circulated to the incident management team and other relevant people in the HPA on 27 June 2008. Data from the HPA/NHS Direct and HPA/QSurveillance systems were provided in a series of regular reports, initially daily and eventually weekly, until the final report on 21 August 2008. Each report included a summary interpretation and more detailed data on diarrhoea, gastroenteritis and vomiting indicators. Methods Surveillance systems HPA/NHS Direct surveillance system NHS Direct is a 24-hour nurse-led telephone helpline that provides health information and advice to the general public. Nurses use a computerised clinical decision support system – the NHS Clinical Assessment System (NHS CAS) – to handle calls. This assessment system uses approximately 200 computerised symptom-based clinical algorithms. Nurses assign the call to the most appropriate algorithm and the patient’s symptoms determine the questions asked and the action to be taken following the call (call outcome), which could be guidance on self-care or they could be referred to their GP or advised to attend a hospital emergency department. No attempt is made to provide a formal diagnosis. Daily NHS Direct data are received by the Real-time Syndromic Surveillance Team, where the number and type of calls received during the previous day are analysed and interpreted. Call proportions are calculated by age group and algorithm against the total number of calls received. HPA/QSurveillance system The HPA/QSurveillance national surveillance system was set up by the University of Nottingham, United Kingdom, and Egton Medical Information Systems (EMIS), a supplier of general practice computer systems, in collaboration with the HPA. It comprises a network of more than 3,500 general practices throughout the UK, covering more than 22 million patients (about 38% of the population [28]). Aggregated data on GP consultations for a range of indicators are automatically uploaded daily from GP practice systems to a central database. Data are routinely reported on a weekly basis; however, daily reporting is possible for specific incidents. Reports are provided at national or regional level (strategic health authority, SHA) and by local health district (primary care trust, PCT). Analysis of surveillance data NHS Direct call proportions for gastrointestinal syndromes (diarrhoea and vomiting) for the East Midlands region in England, where Northampton is situated, were examined during the outbreak (24 June – 18 July 2008) and compared with those for England and Wales. A series of control charts for diarrhoea calls are routinely used to monitor significant rises in the numbers of calls received. Control charts are calculated by assuming that calls follow a Poisson distribution with the total number of calls as an offset: a model is fitted to each region and symptom separately [29]. The model takes into account call variation caused by weekends, public holidays and the time of year – variables that can affect the number of calls received by NHS Direct. A value above the upper limit of the 99.5% confidence interval would be considered to be unusual. The seven-day moving average for diarrhoea calls was also monitored. The number and percentage of calls for diarrhoea in the East Midlands region were presented by call outcome and the number of calls in the Northampton (NN) postcode districts and in particular the number of calls in the NN11 and NN12 postcode districts, which were most affected by the incident. QSurveillance national consultation rates per 100,000 population for diarrhoea (in the age groups under five years, five years and over, and all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with rates for the same period in 2007 (data not presented). Consultation rates by region for 2008 for diarrhoea (all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with those for the East Midlands region. The gastroenteritis indicator includes all cases of diarrhoea and/or vomiting. Consultation rates and standardised incidence ratios (SIRs) – calculated using the UK as the standard population – for diarrhoea, gastroenteritis and vomiting were compared for the UK, Yorkshire and Humberside, East Midlands, Leicestershire, Northamptonshire and Rutland SHA, and Daventry and South Northants PCT, Northamptonshire Heartlands PCT and Northampton PCT. Yorkshire and Humberside was not an affected region but was included as a control. The area supplied by the Pitsford Reservoir included the three PCTs, which were all within the Leicestershire, Northamptonshire and Rutland SHA. The consultation rates and SIRs were compared for the period from week 16 to week 35 of 2008 in order to compare the rates before and after the Cryptosporidium exceedance, which took place in week 26. Estimates of excess numbers of cases of diarrhoea occurring during and following the Cryptosporidium outbreak were made by calculating the mean consultation rate over a period of five weeks before and after the incident (weeks 20–24 and weeks 31–35, respectively). For each of the three PCTs, the calculated mean rate was applied to the PCT population to estimate the number of cases that would be expected each week. The actual consultation rates for diarrhoea for weeks 25 to 30 were used to estimate the number of cases for the PCT population each week. The expected number of cases was subtracted from the estimated number of cases in the PCT population to give the estimated number of excess cases. Results HPA/NHS Direct surveillance system A peak in the number of calls for diarrhoea in the East Midlands was recorded in 25–26 June 2008, the period that coincided with the contamination incident and the associated media coverage (Figure 1). The neighbouring areas of the West Midlands, Yorkshire and the Humber, and East of England showed no increase in the number of calls for diarrhoea. The peak produced a control chart exceedance for calls for diarrhoea on 25 June 2008 (Figure 2), when the proportion of calls exceeded the upper limit of the 99.5% confidence interval. There were further confidence interval exceedances on 26 and 28 June (which were not control chart exceedances). There was no peak in calls for vomiting or control chart exceedance for these calls in the East Midlands. HPA/QSurveillance national surveillance system The East Midlands region had significantly high consultation rates for diarrhoea and gastroenteritis in week 25 (16–22 June), week 26 (23–29 June 2008, when the contamination incident was reported) and in the following four weeks. Within this region. Leicestershire, Northamptonshire and Rutland SHA had slightly raised consultation rates and significant SIRs across weeks 25 to 30 that were not seen in the neighbouring Trent SHA. At PCT level, all three of the PCTs in the area affected by the incident showed increased consultation rates for diarrhoea (Table 1) and gastroenteritis (Table 2) with SIRs significantly above the UK rate in week 26. Daventry and South Northants PCT also had a raised SIR for both indicators in week 25, and although Northamptonshire Heartlands and Northampton PCTs did not have SIRs significantly above that of the UK in week 25, the rise in consultation rates for diarrhoea and gastroenteritis began during week 25. In Northampton PCT, consultations for both diarrhoea and gastroenteritis peaked in the week following the contamination incident, week 27, returning to normal levels by week 30 (Figure 3A and 3B). A similar effect can be seen in Northampton Heartlands PCT. Daventry and South Northants PCT also showed an increase, but appeared to have consistently higher rates for both indicators. This was the area with the smallest population so the rates were more variable than in the other PCTs and we therefore interpreted these results with caution. The consultation rates for vomiting during weeks 25 to 30 in the East Midlands were not unusual at SHA or PCT level (data not presented). Discussion We have demonstrated the sensitivity of syndromic surveillance in detecting this small Cryptosporidium outbreak and the value of the surveillance in being able to describe the extent of its spread. Both the HPA/NHS Direct and HPA/QSurveillance systems showed demonstrable increases in calls and consultations for diarrhoea that were linked to the outbreak. QSurveillance consultations appeared to increase across the PCTs immediately affected but not in the surrounding area. Both the HPA/NHS Direct and HPA/QSurveillance systems showed a clear signal at the time of the incident and we were able to describe the extent of the impact on pre-primary care and primary care services. The HPA/QSurveillance system showed a rise in consultation rates for gastrointestinal symptoms that began the week before the outbreak, consistent with the period when Cryptosporidium was present in the water leaving the Pitsford Reservoir (19–24 June 2008) and with the onset of symptoms in the first outbreak case on 24 June. Although only 33 cases were identified by the outbreak investigation team, of which 23 were confirmed as having the outbreak Cryptosporidium strain, our syndromic surveillance data detected this limited outbreak. Data also suggested a more widespread increase in general gastrointestinal symptoms around the time of the outbreak, with an estimated 422 excess diarrhoea cases; these excess cases represented an increase of about 25% above normally expected levels. It is highly probable that a proportion of these excess cases may have resulted from the increased publicity surrounding the incident – for example, it is likely that media reports contributed to the large peak in calls detected by the HPA/NHS Direct surveillance system on the day the boil water notice was issued, and could also have impacted on the GP consultation rate. It has been previously shown that reporting of mumps cases is sensitive to media coverage, with a rise in clinically reported cases following newspaper reports [30]. A similar mechanism could account for some of the excess GP consultations as cases experiencing gastrointestinal symptoms may have been more likely to consult their GP, whereas in normal circumstances they would have cared for themselves at home. It is also possible that the surveillance shows outbreak-associated cases that did not come to the attention of the outbreak team, perhaps because symptoms were not sufficiently severe to warrant further investigation, or stool samples were not provided for testing. It is interesting to note that there was no demonstrable impact on the number of calls for vomiting (which is not a prominent clinical feature of cryptosporidiosis). Other common community-based pathogens such as norovirus and rotavirus were at low levels, as is normal for that time of year [31]. In this instance, public health authorities had already been alerted to a potential problem by the water company, although the extent of the outbreak was detected by syndromic surveillance. In 2003 the syndromic surveillance systems in the city of New York, United States, were able to detect an increase in diarrhoeal illness following a power outage when there was no other indication of citywide illness [32]. The New York system covers a population of nine million, but does not regularly detect localised outbreaks [33]. It has been shown previously that the HPA/NHS Direct surveillance system would be unlikely to detect a Cryptosporidium outbreak unless call volumes are high (72% chance of detection if nine-tenths of cases called NHS Direct) [29], although the value of syndromic surveillance for such outbreaks has been recognised [34]. The system detected the East Midlands Cryptosporidium outbreak that affected a smaller population than that covered by the New York system. The three PCTs affected have a combined population of around 600,000, of which just over half use GP practices reporting to QSurveillance, yet this syndromic surveillance system was able to describe an increase in consultation rates for diarrhoea and gastroenteritis around the time of the outbreak. Limitations of the data There was extensive media reporting of the incident that may have affected both the HPA/NHS Direct and HPA/QSurveillance systems and contributed to the increase in reported gastrointestinal symptoms around the time of the contamination incident. However, the rise in consultation rates for diarrhoea began before the outbreak had been detected and therefore cannot be attributed to media coverage. The HPA/NHS Direct and HPA/QSurveillance systems monitor general symptoms and so could only monitor the relevant symptoms of diarrhoea and vomiting. They are not able to detect Cryptosporidium cases, as this would require laboratory confirmation of diagnosis, so some of the estimated excess cases could be unconnected with this incident. This outbreak was discovered by other means but both the HPA/NHS Direct and HPA/ QSurveillance systems were able to describe the extent of the disease in the general population and provide reassurance that there was no widespread impact. Compared with other populations, older people and ethnic minorities are less likely to call NHS Direct [29], and although this should not prevent detection of gastrointestinal symptoms as a result of drinking water contamination as this would affect the whole population, this may reduce the signal from the system [35]. With such large surveillance systems, there will be ‘background noise’ in the data, so procedures must be in place to correctly interpret the data and set appropriate thresholds for action. Conclusion To our knowledge, this is the first time that PCT-level data from a syndromic surveillance system, the HPA/ QSurveillance national surveillance system, have been able to show the extent of such a limited outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide PCT-level data and this Cryptosporidium contamination incident demonstrates the potential usefulness of this system.

What is the date of this event?

1498

Value of syndromic surveillance in monitoring a focal waterborne outbreak due to an unusual Cryptosporidium genotype in Northamptonshire

The United Kingdom (UK) has several national syndromic surveillance systems. The Health Protection Agency (HPA)/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data from a national telephone helpline, while the HPA/ QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems. Data from both of these systems were used to monitor a local outbreak of cryptosporidiosis that occurred following Cryptosporidium oocyst contamination of drinking water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, in June 2008. There was a peak in the number of calls to NHS Direct concerning diarrhoea that coincided with the incident. QSurveillance data for the local areas affected by the outbreak showed a significant increase in GP consultations for diarrhoea and gastroenteritis in the week of the incident but there was no increase in consultations for vomiting. A total of 33 clinical cases of cryptosporidiosis were identified in the outbreak investigation, of which 23 were confirmed as infected with the outbreak strain. However, QSurveillance data suggest that there were an estimated 422 excess diarrhoea cases during the outbreak, an increase of about 25% over baseline weekly levels. To our knowledge, this is the first time that data from a syndromic surveillance system, the HPA/QSurveillance national surveillance system, have been able to show the extent of such a small outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide data at local health district (primary care trust) level. The Cryptosporidium contamination incident described demonstrates the potential usefulness of this information, as it is unusual for syndromic surveillance systems to be able to help monitor such a small-scale outbreak. Introduction As syndromic surveillance systems usually capture data already collected for other purposes, and monitor generic symptoms and/or clinically diagnosed disease, they provide information at an earlier stage of illness (compared with laboratory-confirmed diagnoses), so that action can be taken in time to substantially reduce the impact of disease. Some systems, for example, the Royal College of General Practitioners Weekly Returns Service, are now well established, with many years of historical data that allow monitoring of longer-term disease trends [1]. They have the ability to provide early warning of, for example, seasonal rises in influenza and can trigger public health action, such as a recommendation to prescribe antiviral drugs in line with national guidance [2-4]. They can also provide reassurance to incident response teams and the general public that an incident has not caused adverse health effects – for example, following an explosion at the Buncefield oil storage depot in Hemel Hempstead, United Kingdom (UK), in 2005, syndromic surveillance confirmed that there were no unusual rises in community-based morbidity linked to the incident [5]; following the eruption of the Eyjafjallajökull volcano in Iceland in April 2010 similar assurance was given about lack of impact on community morbidity [6]. Health departments are increasingly expected to monitor health effects of natural events such as heat wave or flooding, or implement surveillance – of which syndromic surveillance plays a major role – for mass gatherings such as the Olympics or football World Cup [7-9]. Systems in France, Australia and Taiwan use data from emergency departments [10-12], a Canadian system uses over-the-counter pharmacy sales [13,14], and in the Netherlands data from both syndromic and surrogate data sources, such as employee absence records and prescription medications dispensed by pharmacies, are included in surveillance systems [15,16]. Currently systems based on Internet searches via search engines or on queries submitted to medical websites are being developed [17,18]. In the UK, the HPA/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data collected from the NHS Direct telephone helpline [19], while the HPA/QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems [20]. The HPA Real-time Syndromic Surveillance Team is a small team that coordinates a number of syndromic surveillance systems within the HPA and takes a lead for syndromic surveillance in England [21]. This paper describes the support provided by the team to the local incident management team during a local cryptosporidiosis outbreak and shows the use of syndromic surveillance in monitoring the extent of an outbreak using the HPA/NHS Direct and HPA/QSurveillance national surveillance systems. Cryptosporidiosis Cryptosporidium is a protozoan parasite that can cause an infection in people, cattle and sometimes other animals [22]. Cryptosporidiosis is most common in children aged between one and five years, but it can affect all ages. Those with impaired immune systems are likely to be most seriously affected. Symptoms usually appear between three and 12 days after initial exposure and include watery diarrhoea, stomach pains, dehydration and fever. In its transmissible form, called an oocyst, the parasite is protected by an outer shell, which allows it to survive in the environment for a long time. Transmission occurs most often via the faeco-oral route through person-to-person or animalto-person contact, but people may also be infected by consuming contaminated water or food or by swimming in contaminated water. Although uncommon, the largest outbreaks have occurred following contamination of drinking water [23,24]. Normal chlorine disinfection procedures do not kill the oocysts, so they are removed by filtration and water companies carry out routine monitoring of treated water. Description of the incident On 25 June 2008 the local Health Protection Unit was informed by Anglian Water of an exceedence in the level of Cryptosporidium oocysts found in water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, during 19 to 24 June 2008 [25]. The reservoir supplied a population of more than 250,000 in the Northampton area. A notice advising people in the affected areas to boil all drinking water was issued on 25 June 2008 and public health messages were circulated to local health services and to the general public via the media. Those members of the public who were concerned about health risks associated with the incident were asked to ring NHS Direct for clinical advice [26]. The HPA wrote to local GPs and hospitals asking them to monitor potential patients for signs and symptoms of Cryptosporidium infection and to submit faecal specimens to the local hospital diagnostic laboratory if patients presented with diarrhoea. Samples from 34 patients where Cryptosporidium infection was identified were sent to the UK Cryptosporidium reference unit for typing. On 30 June 2008, the Cryptosporidium oocysts found in the reservoir water were confirmed as being of the rabbit genotype Cryptosporidium cuniculus [27]. Subsequently, a dead rabbit was found in a treated water tank at the water treatment works. The genotype of Cryptosporidium oocysts in the rabbit’s large bowel was indistinguishable from that of the oocysts found in the water [27]. After remediation of the water supply and distribution, the ‘boil water notice’ was lifted on 4 July and the following day the first case of cryptosporidiosis linked to the incident was identified by the reference laboratory (this case was infected with C. cuniculus). During the course of the outbreak (24 June – 18 July 2008, the dates of symptom onset in the first and last case, respectively), 23 cases of cryptosporidiosis were confirmed as being infected with C. cuniculus; one of the 23 was a secondary case. The HPA Real-time Syndromic Surveillance Team provided data in order to aid the response to this incident and the first syndromic surveillance report was circulated to the incident management team and other relevant people in the HPA on 27 June 2008. Data from the HPA/NHS Direct and HPA/QSurveillance systems were provided in a series of regular reports, initially daily and eventually weekly, until the final report on 21 August 2008. Each report included a summary interpretation and more detailed data on diarrhoea, gastroenteritis and vomiting indicators. Methods Surveillance systems HPA/NHS Direct surveillance system NHS Direct is a 24-hour nurse-led telephone helpline that provides health information and advice to the general public. Nurses use a computerised clinical decision support system – the NHS Clinical Assessment System (NHS CAS) – to handle calls. This assessment system uses approximately 200 computerised symptom-based clinical algorithms. Nurses assign the call to the most appropriate algorithm and the patient’s symptoms determine the questions asked and the action to be taken following the call (call outcome), which could be guidance on self-care or they could be referred to their GP or advised to attend a hospital emergency department. No attempt is made to provide a formal diagnosis. Daily NHS Direct data are received by the Real-time Syndromic Surveillance Team, where the number and type of calls received during the previous day are analysed and interpreted. Call proportions are calculated by age group and algorithm against the total number of calls received. HPA/QSurveillance system The HPA/QSurveillance national surveillance system was set up by the University of Nottingham, United Kingdom, and Egton Medical Information Systems (EMIS), a supplier of general practice computer systems, in collaboration with the HPA. It comprises a network of more than 3,500 general practices throughout the UK, covering more than 22 million patients (about 38% of the population [28]). Aggregated data on GP consultations for a range of indicators are automatically uploaded daily from GP practice systems to a central database. Data are routinely reported on a weekly basis; however, daily reporting is possible for specific incidents. Reports are provided at national or regional level (strategic health authority, SHA) and by local health district (primary care trust, PCT). Analysis of surveillance data NHS Direct call proportions for gastrointestinal syndromes (diarrhoea and vomiting) for the East Midlands region in England, where Northampton is situated, were examined during the outbreak (24 June – 18 July 2008) and compared with those for England and Wales. A series of control charts for diarrhoea calls are routinely used to monitor significant rises in the numbers of calls received. Control charts are calculated by assuming that calls follow a Poisson distribution with the total number of calls as an offset: a model is fitted to each region and symptom separately [29]. The model takes into account call variation caused by weekends, public holidays and the time of year – variables that can affect the number of calls received by NHS Direct. A value above the upper limit of the 99.5% confidence interval would be considered to be unusual. The seven-day moving average for diarrhoea calls was also monitored. The number and percentage of calls for diarrhoea in the East Midlands region were presented by call outcome and the number of calls in the Northampton (NN) postcode districts and in particular the number of calls in the NN11 and NN12 postcode districts, which were most affected by the incident. QSurveillance national consultation rates per 100,000 population for diarrhoea (in the age groups under five years, five years and over, and all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with rates for the same period in 2007 (data not presented). Consultation rates by region for 2008 for diarrhoea (all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with those for the East Midlands region. The gastroenteritis indicator includes all cases of diarrhoea and/or vomiting. Consultation rates and standardised incidence ratios (SIRs) – calculated using the UK as the standard population – for diarrhoea, gastroenteritis and vomiting were compared for the UK, Yorkshire and Humberside, East Midlands, Leicestershire, Northamptonshire and Rutland SHA, and Daventry and South Northants PCT, Northamptonshire Heartlands PCT and Northampton PCT. Yorkshire and Humberside was not an affected region but was included as a control. The area supplied by the Pitsford Reservoir included the three PCTs, which were all within the Leicestershire, Northamptonshire and Rutland SHA. The consultation rates and SIRs were compared for the period from week 16 to week 35 of 2008 in order to compare the rates before and after the Cryptosporidium exceedance, which took place in week 26. Estimates of excess numbers of cases of diarrhoea occurring during and following the Cryptosporidium outbreak were made by calculating the mean consultation rate over a period of five weeks before and after the incident (weeks 20–24 and weeks 31–35, respectively). For each of the three PCTs, the calculated mean rate was applied to the PCT population to estimate the number of cases that would be expected each week. The actual consultation rates for diarrhoea for weeks 25 to 30 were used to estimate the number of cases for the PCT population each week. The expected number of cases was subtracted from the estimated number of cases in the PCT population to give the estimated number of excess cases. Results HPA/NHS Direct surveillance system A peak in the number of calls for diarrhoea in the East Midlands was recorded in 25–26 June 2008, the period that coincided with the contamination incident and the associated media coverage (Figure 1). The neighbouring areas of the West Midlands, Yorkshire and the Humber, and East of England showed no increase in the number of calls for diarrhoea. The peak produced a control chart exceedance for calls for diarrhoea on 25 June 2008 (Figure 2), when the proportion of calls exceeded the upper limit of the 99.5% confidence interval. There were further confidence interval exceedances on 26 and 28 June (which were not control chart exceedances). There was no peak in calls for vomiting or control chart exceedance for these calls in the East Midlands. HPA/QSurveillance national surveillance system The East Midlands region had significantly high consultation rates for diarrhoea and gastroenteritis in week 25 (16–22 June), week 26 (23–29 June 2008, when the contamination incident was reported) and in the following four weeks. Within this region. Leicestershire, Northamptonshire and Rutland SHA had slightly raised consultation rates and significant SIRs across weeks 25 to 30 that were not seen in the neighbouring Trent SHA. At PCT level, all three of the PCTs in the area affected by the incident showed increased consultation rates for diarrhoea (Table 1) and gastroenteritis (Table 2) with SIRs significantly above the UK rate in week 26. Daventry and South Northants PCT also had a raised SIR for both indicators in week 25, and although Northamptonshire Heartlands and Northampton PCTs did not have SIRs significantly above that of the UK in week 25, the rise in consultation rates for diarrhoea and gastroenteritis began during week 25. In Northampton PCT, consultations for both diarrhoea and gastroenteritis peaked in the week following the contamination incident, week 27, returning to normal levels by week 30 (Figure 3A and 3B). A similar effect can be seen in Northampton Heartlands PCT. Daventry and South Northants PCT also showed an increase, but appeared to have consistently higher rates for both indicators. This was the area with the smallest population so the rates were more variable than in the other PCTs and we therefore interpreted these results with caution. The consultation rates for vomiting during weeks 25 to 30 in the East Midlands were not unusual at SHA or PCT level (data not presented). Discussion We have demonstrated the sensitivity of syndromic surveillance in detecting this small Cryptosporidium outbreak and the value of the surveillance in being able to describe the extent of its spread. Both the HPA/NHS Direct and HPA/QSurveillance systems showed demonstrable increases in calls and consultations for diarrhoea that were linked to the outbreak. QSurveillance consultations appeared to increase across the PCTs immediately affected but not in the surrounding area. Both the HPA/NHS Direct and HPA/QSurveillance systems showed a clear signal at the time of the incident and we were able to describe the extent of the impact on pre-primary care and primary care services. The HPA/QSurveillance system showed a rise in consultation rates for gastrointestinal symptoms that began the week before the outbreak, consistent with the period when Cryptosporidium was present in the water leaving the Pitsford Reservoir (19–24 June 2008) and with the onset of symptoms in the first outbreak case on 24 June. Although only 33 cases were identified by the outbreak investigation team, of which 23 were confirmed as having the outbreak Cryptosporidium strain, our syndromic surveillance data detected this limited outbreak. Data also suggested a more widespread increase in general gastrointestinal symptoms around the time of the outbreak, with an estimated 422 excess diarrhoea cases; these excess cases represented an increase of about 25% above normally expected levels. It is highly probable that a proportion of these excess cases may have resulted from the increased publicity surrounding the incident – for example, it is likely that media reports contributed to the large peak in calls detected by the HPA/NHS Direct surveillance system on the day the boil water notice was issued, and could also have impacted on the GP consultation rate. It has been previously shown that reporting of mumps cases is sensitive to media coverage, with a rise in clinically reported cases following newspaper reports [30]. A similar mechanism could account for some of the excess GP consultations as cases experiencing gastrointestinal symptoms may have been more likely to consult their GP, whereas in normal circumstances they would have cared for themselves at home. It is also possible that the surveillance shows outbreak-associated cases that did not come to the attention of the outbreak team, perhaps because symptoms were not sufficiently severe to warrant further investigation, or stool samples were not provided for testing. It is interesting to note that there was no demonstrable impact on the number of calls for vomiting (which is not a prominent clinical feature of cryptosporidiosis). Other common community-based pathogens such as norovirus and rotavirus were at low levels, as is normal for that time of year [31]. In this instance, public health authorities had already been alerted to a potential problem by the water company, although the extent of the outbreak was detected by syndromic surveillance. In 2003 the syndromic surveillance systems in the city of New York, United States, were able to detect an increase in diarrhoeal illness following a power outage when there was no other indication of citywide illness [32]. The New York system covers a population of nine million, but does not regularly detect localised outbreaks [33]. It has been shown previously that the HPA/NHS Direct surveillance system would be unlikely to detect a Cryptosporidium outbreak unless call volumes are high (72% chance of detection if nine-tenths of cases called NHS Direct) [29], although the value of syndromic surveillance for such outbreaks has been recognised [34]. The system detected the East Midlands Cryptosporidium outbreak that affected a smaller population than that covered by the New York system. The three PCTs affected have a combined population of around 600,000, of which just over half use GP practices reporting to QSurveillance, yet this syndromic surveillance system was able to describe an increase in consultation rates for diarrhoea and gastroenteritis around the time of the outbreak. Limitations of the data There was extensive media reporting of the incident that may have affected both the HPA/NHS Direct and HPA/QSurveillance systems and contributed to the increase in reported gastrointestinal symptoms around the time of the contamination incident. However, the rise in consultation rates for diarrhoea began before the outbreak had been detected and therefore cannot be attributed to media coverage. The HPA/NHS Direct and HPA/QSurveillance systems monitor general symptoms and so could only monitor the relevant symptoms of diarrhoea and vomiting. They are not able to detect Cryptosporidium cases, as this would require laboratory confirmation of diagnosis, so some of the estimated excess cases could be unconnected with this incident. This outbreak was discovered by other means but both the HPA/NHS Direct and HPA/ QSurveillance systems were able to describe the extent of the disease in the general population and provide reassurance that there was no widespread impact. Compared with other populations, older people and ethnic minorities are less likely to call NHS Direct [29], and although this should not prevent detection of gastrointestinal symptoms as a result of drinking water contamination as this would affect the whole population, this may reduce the signal from the system [35]. With such large surveillance systems, there will be ‘background noise’ in the data, so procedures must be in place to correctly interpret the data and set appropriate thresholds for action. Conclusion To our knowledge, this is the first time that PCT-level data from a syndromic surveillance system, the HPA/ QSurveillance national surveillance system, have been able to show the extent of such a limited outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide PCT-level data and this Cryptosporidium contamination incident demonstrates the potential usefulness of this system.

How long was the event?

1499

Value of syndromic surveillance in monitoring a focal waterborne outbreak due to an unusual Cryptosporidium genotype in Northamptonshire

The United Kingdom (UK) has several national syndromic surveillance systems. The Health Protection Agency (HPA)/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data from a national telephone helpline, while the HPA/ QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems. Data from both of these systems were used to monitor a local outbreak of cryptosporidiosis that occurred following Cryptosporidium oocyst contamination of drinking water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, in June 2008. There was a peak in the number of calls to NHS Direct concerning diarrhoea that coincided with the incident. QSurveillance data for the local areas affected by the outbreak showed a significant increase in GP consultations for diarrhoea and gastroenteritis in the week of the incident but there was no increase in consultations for vomiting. A total of 33 clinical cases of cryptosporidiosis were identified in the outbreak investigation, of which 23 were confirmed as infected with the outbreak strain. However, QSurveillance data suggest that there were an estimated 422 excess diarrhoea cases during the outbreak, an increase of about 25% over baseline weekly levels. To our knowledge, this is the first time that data from a syndromic surveillance system, the HPA/QSurveillance national surveillance system, have been able to show the extent of such a small outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide data at local health district (primary care trust) level. The Cryptosporidium contamination incident described demonstrates the potential usefulness of this information, as it is unusual for syndromic surveillance systems to be able to help monitor such a small-scale outbreak. Introduction As syndromic surveillance systems usually capture data already collected for other purposes, and monitor generic symptoms and/or clinically diagnosed disease, they provide information at an earlier stage of illness (compared with laboratory-confirmed diagnoses), so that action can be taken in time to substantially reduce the impact of disease. Some systems, for example, the Royal College of General Practitioners Weekly Returns Service, are now well established, with many years of historical data that allow monitoring of longer-term disease trends [1]. They have the ability to provide early warning of, for example, seasonal rises in influenza and can trigger public health action, such as a recommendation to prescribe antiviral drugs in line with national guidance [2-4]. They can also provide reassurance to incident response teams and the general public that an incident has not caused adverse health effects – for example, following an explosion at the Buncefield oil storage depot in Hemel Hempstead, United Kingdom (UK), in 2005, syndromic surveillance confirmed that there were no unusual rises in community-based morbidity linked to the incident [5]; following the eruption of the Eyjafjallajökull volcano in Iceland in April 2010 similar assurance was given about lack of impact on community morbidity [6]. Health departments are increasingly expected to monitor health effects of natural events such as heat wave or flooding, or implement surveillance – of which syndromic surveillance plays a major role – for mass gatherings such as the Olympics or football World Cup [7-9]. Systems in France, Australia and Taiwan use data from emergency departments [10-12], a Canadian system uses over-the-counter pharmacy sales [13,14], and in the Netherlands data from both syndromic and surrogate data sources, such as employee absence records and prescription medications dispensed by pharmacies, are included in surveillance systems [15,16]. Currently systems based on Internet searches via search engines or on queries submitted to medical websites are being developed [17,18]. In the UK, the HPA/NHS Direct syndromic surveillance system uses pre-diagnostic syndromic data collected from the NHS Direct telephone helpline [19], while the HPA/QSurveillance national surveillance system uses clinical diagnosis data extracted from general practitioner (GP)-based clinical information systems [20]. The HPA Real-time Syndromic Surveillance Team is a small team that coordinates a number of syndromic surveillance systems within the HPA and takes a lead for syndromic surveillance in England [21]. This paper describes the support provided by the team to the local incident management team during a local cryptosporidiosis outbreak and shows the use of syndromic surveillance in monitoring the extent of an outbreak using the HPA/NHS Direct and HPA/QSurveillance national surveillance systems. Cryptosporidiosis Cryptosporidium is a protozoan parasite that can cause an infection in people, cattle and sometimes other animals [22]. Cryptosporidiosis is most common in children aged between one and five years, but it can affect all ages. Those with impaired immune systems are likely to be most seriously affected. Symptoms usually appear between three and 12 days after initial exposure and include watery diarrhoea, stomach pains, dehydration and fever. In its transmissible form, called an oocyst, the parasite is protected by an outer shell, which allows it to survive in the environment for a long time. Transmission occurs most often via the faeco-oral route through person-to-person or animalto-person contact, but people may also be infected by consuming contaminated water or food or by swimming in contaminated water. Although uncommon, the largest outbreaks have occurred following contamination of drinking water [23,24]. Normal chlorine disinfection procedures do not kill the oocysts, so they are removed by filtration and water companies carry out routine monitoring of treated water. Description of the incident On 25 June 2008 the local Health Protection Unit was informed by Anglian Water of an exceedence in the level of Cryptosporidium oocysts found in water supplied from the Pitsford Reservoir in Northamptonshire, United Kingdom, during 19 to 24 June 2008 [25]. The reservoir supplied a population of more than 250,000 in the Northampton area. A notice advising people in the affected areas to boil all drinking water was issued on 25 June 2008 and public health messages were circulated to local health services and to the general public via the media. Those members of the public who were concerned about health risks associated with the incident were asked to ring NHS Direct for clinical advice [26]. The HPA wrote to local GPs and hospitals asking them to monitor potential patients for signs and symptoms of Cryptosporidium infection and to submit faecal specimens to the local hospital diagnostic laboratory if patients presented with diarrhoea. Samples from 34 patients where Cryptosporidium infection was identified were sent to the UK Cryptosporidium reference unit for typing. On 30 June 2008, the Cryptosporidium oocysts found in the reservoir water were confirmed as being of the rabbit genotype Cryptosporidium cuniculus [27]. Subsequently, a dead rabbit was found in a treated water tank at the water treatment works. The genotype of Cryptosporidium oocysts in the rabbit’s large bowel was indistinguishable from that of the oocysts found in the water [27]. After remediation of the water supply and distribution, the ‘boil water notice’ was lifted on 4 July and the following day the first case of cryptosporidiosis linked to the incident was identified by the reference laboratory (this case was infected with C. cuniculus). During the course of the outbreak (24 June – 18 July 2008, the dates of symptom onset in the first and last case, respectively), 23 cases of cryptosporidiosis were confirmed as being infected with C. cuniculus; one of the 23 was a secondary case. The HPA Real-time Syndromic Surveillance Team provided data in order to aid the response to this incident and the first syndromic surveillance report was circulated to the incident management team and other relevant people in the HPA on 27 June 2008. Data from the HPA/NHS Direct and HPA/QSurveillance systems were provided in a series of regular reports, initially daily and eventually weekly, until the final report on 21 August 2008. Each report included a summary interpretation and more detailed data on diarrhoea, gastroenteritis and vomiting indicators. Methods Surveillance systems HPA/NHS Direct surveillance system NHS Direct is a 24-hour nurse-led telephone helpline that provides health information and advice to the general public. Nurses use a computerised clinical decision support system – the NHS Clinical Assessment System (NHS CAS) – to handle calls. This assessment system uses approximately 200 computerised symptom-based clinical algorithms. Nurses assign the call to the most appropriate algorithm and the patient’s symptoms determine the questions asked and the action to be taken following the call (call outcome), which could be guidance on self-care or they could be referred to their GP or advised to attend a hospital emergency department. No attempt is made to provide a formal diagnosis. Daily NHS Direct data are received by the Real-time Syndromic Surveillance Team, where the number and type of calls received during the previous day are analysed and interpreted. Call proportions are calculated by age group and algorithm against the total number of calls received. HPA/QSurveillance system The HPA/QSurveillance national surveillance system was set up by the University of Nottingham, United Kingdom, and Egton Medical Information Systems (EMIS), a supplier of general practice computer systems, in collaboration with the HPA. It comprises a network of more than 3,500 general practices throughout the UK, covering more than 22 million patients (about 38% of the population [28]). Aggregated data on GP consultations for a range of indicators are automatically uploaded daily from GP practice systems to a central database. Data are routinely reported on a weekly basis; however, daily reporting is possible for specific incidents. Reports are provided at national or regional level (strategic health authority, SHA) and by local health district (primary care trust, PCT). Analysis of surveillance data NHS Direct call proportions for gastrointestinal syndromes (diarrhoea and vomiting) for the East Midlands region in England, where Northampton is situated, were examined during the outbreak (24 June – 18 July 2008) and compared with those for England and Wales. A series of control charts for diarrhoea calls are routinely used to monitor significant rises in the numbers of calls received. Control charts are calculated by assuming that calls follow a Poisson distribution with the total number of calls as an offset: a model is fitted to each region and symptom separately [29]. The model takes into account call variation caused by weekends, public holidays and the time of year – variables that can affect the number of calls received by NHS Direct. A value above the upper limit of the 99.5% confidence interval would be considered to be unusual. The seven-day moving average for diarrhoea calls was also monitored. The number and percentage of calls for diarrhoea in the East Midlands region were presented by call outcome and the number of calls in the Northampton (NN) postcode districts and in particular the number of calls in the NN11 and NN12 postcode districts, which were most affected by the incident. QSurveillance national consultation rates per 100,000 population for diarrhoea (in the age groups under five years, five years and over, and all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with rates for the same period in 2007 (data not presented). Consultation rates by region for 2008 for diarrhoea (all ages), gastroenteritis (all ages) and vomiting (all ages) were compared with those for the East Midlands region. The gastroenteritis indicator includes all cases of diarrhoea and/or vomiting. Consultation rates and standardised incidence ratios (SIRs) – calculated using the UK as the standard population – for diarrhoea, gastroenteritis and vomiting were compared for the UK, Yorkshire and Humberside, East Midlands, Leicestershire, Northamptonshire and Rutland SHA, and Daventry and South Northants PCT, Northamptonshire Heartlands PCT and Northampton PCT. Yorkshire and Humberside was not an affected region but was included as a control. The area supplied by the Pitsford Reservoir included the three PCTs, which were all within the Leicestershire, Northamptonshire and Rutland SHA. The consultation rates and SIRs were compared for the period from week 16 to week 35 of 2008 in order to compare the rates before and after the Cryptosporidium exceedance, which took place in week 26. Estimates of excess numbers of cases of diarrhoea occurring during and following the Cryptosporidium outbreak were made by calculating the mean consultation rate over a period of five weeks before and after the incident (weeks 20–24 and weeks 31–35, respectively). For each of the three PCTs, the calculated mean rate was applied to the PCT population to estimate the number of cases that would be expected each week. The actual consultation rates for diarrhoea for weeks 25 to 30 were used to estimate the number of cases for the PCT population each week. The expected number of cases was subtracted from the estimated number of cases in the PCT population to give the estimated number of excess cases. Results HPA/NHS Direct surveillance system A peak in the number of calls for diarrhoea in the East Midlands was recorded in 25–26 June 2008, the period that coincided with the contamination incident and the associated media coverage (Figure 1). The neighbouring areas of the West Midlands, Yorkshire and the Humber, and East of England showed no increase in the number of calls for diarrhoea. The peak produced a control chart exceedance for calls for diarrhoea on 25 June 2008 (Figure 2), when the proportion of calls exceeded the upper limit of the 99.5% confidence interval. There were further confidence interval exceedances on 26 and 28 June (which were not control chart exceedances). There was no peak in calls for vomiting or control chart exceedance for these calls in the East Midlands. HPA/QSurveillance national surveillance system The East Midlands region had significantly high consultation rates for diarrhoea and gastroenteritis in week 25 (16–22 June), week 26 (23–29 June 2008, when the contamination incident was reported) and in the following four weeks. Within this region. Leicestershire, Northamptonshire and Rutland SHA had slightly raised consultation rates and significant SIRs across weeks 25 to 30 that were not seen in the neighbouring Trent SHA. At PCT level, all three of the PCTs in the area affected by the incident showed increased consultation rates for diarrhoea (Table 1) and gastroenteritis (Table 2) with SIRs significantly above the UK rate in week 26. Daventry and South Northants PCT also had a raised SIR for both indicators in week 25, and although Northamptonshire Heartlands and Northampton PCTs did not have SIRs significantly above that of the UK in week 25, the rise in consultation rates for diarrhoea and gastroenteritis began during week 25. In Northampton PCT, consultations for both diarrhoea and gastroenteritis peaked in the week following the contamination incident, week 27, returning to normal levels by week 30 (Figure 3A and 3B). A similar effect can be seen in Northampton Heartlands PCT. Daventry and South Northants PCT also showed an increase, but appeared to have consistently higher rates for both indicators. This was the area with the smallest population so the rates were more variable than in the other PCTs and we therefore interpreted these results with caution. The consultation rates for vomiting during weeks 25 to 30 in the East Midlands were not unusual at SHA or PCT level (data not presented). Discussion We have demonstrated the sensitivity of syndromic surveillance in detecting this small Cryptosporidium outbreak and the value of the surveillance in being able to describe the extent of its spread. Both the HPA/NHS Direct and HPA/QSurveillance systems showed demonstrable increases in calls and consultations for diarrhoea that were linked to the outbreak. QSurveillance consultations appeared to increase across the PCTs immediately affected but not in the surrounding area. Both the HPA/NHS Direct and HPA/QSurveillance systems showed a clear signal at the time of the incident and we were able to describe the extent of the impact on pre-primary care and primary care services. The HPA/QSurveillance system showed a rise in consultation rates for gastrointestinal symptoms that began the week before the outbreak, consistent with the period when Cryptosporidium was present in the water leaving the Pitsford Reservoir (19–24 June 2008) and with the onset of symptoms in the first outbreak case on 24 June. Although only 33 cases were identified by the outbreak investigation team, of which 23 were confirmed as having the outbreak Cryptosporidium strain, our syndromic surveillance data detected this limited outbreak. Data also suggested a more widespread increase in general gastrointestinal symptoms around the time of the outbreak, with an estimated 422 excess diarrhoea cases; these excess cases represented an increase of about 25% above normally expected levels. It is highly probable that a proportion of these excess cases may have resulted from the increased publicity surrounding the incident – for example, it is likely that media reports contributed to the large peak in calls detected by the HPA/NHS Direct surveillance system on the day the boil water notice was issued, and could also have impacted on the GP consultation rate. It has been previously shown that reporting of mumps cases is sensitive to media coverage, with a rise in clinically reported cases following newspaper reports [30]. A similar mechanism could account for some of the excess GP consultations as cases experiencing gastrointestinal symptoms may have been more likely to consult their GP, whereas in normal circumstances they would have cared for themselves at home. It is also possible that the surveillance shows outbreak-associated cases that did not come to the attention of the outbreak team, perhaps because symptoms were not sufficiently severe to warrant further investigation, or stool samples were not provided for testing. It is interesting to note that there was no demonstrable impact on the number of calls for vomiting (which is not a prominent clinical feature of cryptosporidiosis). Other common community-based pathogens such as norovirus and rotavirus were at low levels, as is normal for that time of year [31]. In this instance, public health authorities had already been alerted to a potential problem by the water company, although the extent of the outbreak was detected by syndromic surveillance. In 2003 the syndromic surveillance systems in the city of New York, United States, were able to detect an increase in diarrhoeal illness following a power outage when there was no other indication of citywide illness [32]. The New York system covers a population of nine million, but does not regularly detect localised outbreaks [33]. It has been shown previously that the HPA/NHS Direct surveillance system would be unlikely to detect a Cryptosporidium outbreak unless call volumes are high (72% chance of detection if nine-tenths of cases called NHS Direct) [29], although the value of syndromic surveillance for such outbreaks has been recognised [34]. The system detected the East Midlands Cryptosporidium outbreak that affected a smaller population than that covered by the New York system. The three PCTs affected have a combined population of around 600,000, of which just over half use GP practices reporting to QSurveillance, yet this syndromic surveillance system was able to describe an increase in consultation rates for diarrhoea and gastroenteritis around the time of the outbreak. Limitations of the data There was extensive media reporting of the incident that may have affected both the HPA/NHS Direct and HPA/QSurveillance systems and contributed to the increase in reported gastrointestinal symptoms around the time of the contamination incident. However, the rise in consultation rates for diarrhoea began before the outbreak had been detected and therefore cannot be attributed to media coverage. The HPA/NHS Direct and HPA/QSurveillance systems monitor general symptoms and so could only monitor the relevant symptoms of diarrhoea and vomiting. They are not able to detect Cryptosporidium cases, as this would require laboratory confirmation of diagnosis, so some of the estimated excess cases could be unconnected with this incident. This outbreak was discovered by other means but both the HPA/NHS Direct and HPA/ QSurveillance systems were able to describe the extent of the disease in the general population and provide reassurance that there was no widespread impact. Compared with other populations, older people and ethnic minorities are less likely to call NHS Direct [29], and although this should not prevent detection of gastrointestinal symptoms as a result of drinking water contamination as this would affect the whole population, this may reduce the signal from the system [35]. With such large surveillance systems, there will be ‘background noise’ in the data, so procedures must be in place to correctly interpret the data and set appropriate thresholds for action. Conclusion To our knowledge, this is the first time that PCT-level data from a syndromic surveillance system, the HPA/ QSurveillance national surveillance system, have been able to show the extent of such a limited outbreak at a local level. QSurveillance, which covers about 38% of the UK population, is currently the only GP database that is able to provide PCT-level data and this Cryptosporidium contamination incident demonstrates the potential usefulness of this system.

How long did the event last?