Estimating the primary etiologic agents in recreational freshwaters impacted by human sources of faecal contamination

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Estimating the primary etiologic agents in recreational freshwaters impacted by human sources of faecal contamination5 Jeffrey A. Soller a,*, Timothy Bartrand b, Nicholas J. Ashbolt c, John Ravenscroft d, Timothy J. Wade e a

Soller Environmental, LLC, 3022 King St, Berkeley, CA 94703, USA Clancy Environmental Consultants, PO Box 314, St. Albans, VT 05478, USA c U.S. Environmental Protection Agency, Office of Research and Development, 26 West Martin Luther King Drive, Cincinnati, OH 45268, USA d U.S. Environmental Protection Agency, Office of Water, 1200 Pennsylvania Ave, NW 4304T, Washington DC 20460, USA e U.S. Environmental Protection Agency, Office of Research and Development, MD 58C, Research Triangle Park, NC 27711, USA b

article info

abstract

Article history:

Epidemiology studies of recreational waters have demonstrated that swimmers exposed to

Received 2 April 2010

faecally-contaminated recreational waters are at risk of excess gastrointestinal illness.

Received in revised form

Epidemiology studies provide valuable information on the nature and extent of health

9 July 2010

effects, the magnitude of risks, and how these risks are modified or associated with levels

Accepted 20 July 2010

of faecal contamination and other measures of pollution. However, such studies have not

Available online 29 July 2010

provided information about the specific microbial agents that are responsible for the observed illnesses in swimmers. The objective of this work was to understand more fully

Keywords:

the reported epidemiologic results from studies conducted on the Great Lakes in the US

Water epidemiology

during 2003 and 2004 by identifying pathogens that could have caused the observed

Quantitative microbial risk

illnesses in those studies. We used a Quantitative Microbial Risk Assessment (QMRA)

assessment

approach to estimate the likelihood of pathogen-induced adverse health effects. The

Recreational water

reference pathogens used for this analysis were Norovirus, rotavirus, adenovirus, Cryptosporidium spp., Giardia lamblia, Campylobacter jejuni, Salmonella enterica, and Escherichia coli O157:H7. Two QMRA-based approaches were used to estimate the pathogen combinations that would be consistent with observed illness rates: in the first, swimming-associated gastrointestinal (GI) illnesses were assumed to occur in the same proportion as known illnesses in the US due to all non-foodborne sources, and in the second, pathogens were assumed to occur in the recreational waters in the same proportion as they occur in disinfected secondary effluent. The results indicate that human enteric viruses and in particular, Norovirus could have caused the vast majority of the observed swimmingassociated GI illnesses during the 2003/2004 water epidemiology studies. Evaluation of the time-to-onset of illness strongly supports the principal finding and sensitivity analyses support the overall trends of the analyses even given their substantial uncertainties. ª 2010 Elsevier Ltd. All rights reserved.

5

The views expressed in this article are those of the authors and do not necessarily reflect the views or policies of the U.S. Environmental Protection Agency. * Corresponding author. Tel.: þ1 510 847 0474. E-mail address: [email protected] (J.A. Soller). 0043-1354/$ e see front matter ª 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2010.07.064

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

1.

Introduction

Epidemiology studies on recreational waters have demonstrated that swimmers exposed to faecally-contaminated recreational waters are at risk of excess gastrointestinal (GI) illness (Pru¨ss, 1998; Wade et al., 2003; Zmirou et al., 2003). Moreover, studies conducted by the US EPA and others have further demonstrated exposureeresponse relationships between faecal indicator organisms such as enterococci and E. coli and swimming-associated gastrointestinal illness (Cabelli et al., 1982; Dufour, 1984; Fleisher et al., 1996; Haile et al., 1999; Kay et al., 1994; Wade et al., 2006), and comprehensive literature reviews have concluded general support for these associations in over 20 epidemiology studies conducted around the world (Pru¨ss, 1998; Wade et al., 2003; Zmirou et al., 2003). However, little is known about the specific microbial agents that are responsible for the observed illnesses in swimmers. Nearly 30 years ago, Cabelli (1983) stated: “An intensive program should be initiated towards establishing the etiology of the gastroenteritis observed in these studies and developing methods for quantifying the agent(s) in environmental waters.” While several studies have attempted to collect biologic specimens (blood or stool) as part of epidemiologic research at beach sites, to date these efforts have been largely unsuccessful in identifying the agents responsible for the observed increase in GI symptoms among swimmers (Jones et al., 1991). Information on the etiology of illnesses observed during epidemiology studies is important for several purposes including understanding and managing risks (Ashbolt et al., 2010), developing appropriately protective public health regulations, and potentially extending the observed health relationships to other waters whose characteristics and/or sources are not necessarily covered by the epidemiologic studies (Schoen and Ashbolt, 2010; Soller et al., 2010). We used a Quantitative Microbial Risk Assessment (QMRA) paradigm to understand more fully the reported epidemiologic results from the US EPA National Epidemiological and Environmental Assessment of Recreational (NEEAR) water studies conducted on the Great Lakes in the US during 2003 and 2004 (Wade et al., 2006, 2008) and suggest which pathogens represent the observed illnesses in those studies. US EPA specifically selected the Great Lakes sites because they were thought to be impacted by sewage treatment plant effluent. Beaches were located in proximity to point-source tributaries that received treated wastewater from communities with populations of at least 38,000 and with flow rates of over 10 million gallons per day. These sewage treatment plants provided secondary treatment as well as disinfection with chlorine or ultraviolet radiation during the summer (Wade et al., 2008). QMRA is a process that estimates the likelihood of adverse human health effects that can occur following exposure to pathogens (ILSI, 1996, 2000). This work is one component of a more comprehensive US EPA effort to conduct QMRA on waters impacted by a variety of faecal contamination sources to help facilitate Ambient Water Quality Criteria development and/or implementation for recreational waters (U.S. EPA, 1986, 2007). Quantitative methods to characterize human health risks associated with exposure to pathogens were first published in the 1970s (Dudley et al., 1976; Fuhs, 1975) and have

4737

proliferated in the literature since that time addressing exposures in diverse media including water (Crabtree et al., 1997; Gerba et al., 1996; Haas, 1983; Mena et al., 2003; Regli et al., 1991; Rose et al., 1991; Teunis et al., 1997), food (Buchanan et al., 1998, 2000; Farber et al., 1996), and other media (Brooks et al., 2005; Eisenberg et al., 2004, 2008; Riley et al., 2003). Surprisingly, few QMRA studies have been published that specifically address risks in recreational waters (Ashbolt et al., 1997; Rose et al., 1987; Roser et al., 2007; Schoen and Ashbolt, 2010; Soller et al., 2003, 2006, 2010). Here, we employ QMRA to estimate feasible concentrations of the pathogens of public health concern in a humanimpacted recreational water that would be consistent with the observed NEEAR GI illnesses reported (Wade et al., 2006, 2008). Specifically, using the observed rates of swimming-associated GI illness and presumed faecal contamination source to the waterbodies in the NEEAR studies, we identify the likely etiologic agents causing GI illness. This work is intended to enhance interpretation of results of the epidemiology study and possible extension of the findings to sites for which epidemiologic studies are not planned.

2.

Methods

The analysis entailed matching illness rates observed during the course of epidemiologic studies with predicted illness rates computed using estimated pathogen densities consistent with municipal wastewater/human faecal pollution sources. The analyses began with the faecal indicator densities (Enterococcus qPCR calibrated cell equivalents, CCE) observed during the epidemiologic investigations on the Great Lakes (Wade et al., 2006, 2008). Using the published relationship between qPCR signal and GI illness (Wade et al., 2008), the faecal indicator levels were used to estimate the expected value for the rate of swimming-associated illness for each study day. Additional data were then used to estimate reference pathogen densities in the recreational water consistent with the observed rate of swimming-associated illness on each of the study days. Two complementary approaches were employed to derive those concentrations: 1) a health-based approach in which illnesses are assumed to occur in swimmers in the same proportion as reported illnesses occur in the US due to all non-foodborne sources and, 2) a publicallyowned treatment works [POTW] effluent-based approach in which reference pathogens were assumed to occur in the recreational waters in the same proportion as they are reported to occur in disinfected secondary effluent.

2.1. Analysis of faecal indicator data from the NEEAR studies These data comprised 1600 observations for 78 individual beach exposure days. These data were reduced as reported in the literature (Wade et al., 2008) resulting in a total of 78 daily average mean Enterococcus qPCR values. The swimmingassociated GI illness rate among all swimmers as a function of daily average log10 Enterococcus qPCR CCE was based on a linear probability regression model that adjusted for

4738

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

important confounding factors (Wade et al., 2008). Using this regression model, the expected value of the swimming-associated illness rates for each of the study days was computed.

2.2.

Reference pathogens

The reference pathogens used were Norovirus, rotavirus, adenovirus, Cryptosporidium spp., Giardia lamblia, Campylobacter jejuni, Salmonella enterica, and E. coli O157:H7. Together these eight pathogens have been estimated to make up approximately 97% of all non-foodborne illnesses from known (potentially waterborne) pathogens in the US (calculated based on data from Mead et al., 1999), are considered to represent the fate and transport of other pathogens potentially of concern from the waterborne route of exposure, and have corresponding doseeresponse relationships in the peer reviewed literature. Use of reference pathogens is an accepted practice in the field of QMRA (Roser et al., 2007; Soller et al., 2003, 2006; WHO, 2004).

2.3.

QMRA methods

The estimated swimming-associated illness rates for each of the NEEAR study beach days were used to compute feasible concentrations of the reference pathogens that could have resulted in the observed illness rates among swimmers. It was assumed that the probability of infection from any of the pathogens was independent from others potentially present: co-infection from more than one pathogen was not considered. Additional data describing the mixture of pathogens in the recreational water were necessary. The two approaches used to estimate pathogen densities consistent with reported illness rates are described below.

2.3.1.

Health-based QMRA approach

In the health-based approach, swimming-associated GI illnesses were assumed to occur in the same proportion as projected illnesses from known pathogen occurrence in the US (Mead et al., 1999). Since foodborne sources comprise a large portion of illnesses from known pathogens in the US (Mead et al., 1999), illnesses from foodborne sources were excluded from this analysis to estimate illnesses that potentially could be waterborne. Other potentially important sources, such as spas, swimming pools, interactive fountains are responsible for substantially fewer illnesses in the US than foodborne sources, and therefore, were not excluded from this analysis (CDC, 2006). For human faecally-impacted waters this approach is reasonable if a primary source of contamination is untreated or poorly treated sewage, leaky septic systems, or infected individuals in the recreational waterbody (Ashbolt et al., 2010; Gerba, 2000; Schets et al., 2004). Densities of each of the reference pathogens were computed for each of the study days. Daily densities were then combined to produce averages for each of the reference pathogens across all study days. To accomplish these calculations, the information required included: 1) Percent of the swimming-associated illnesses attributable to each reference pathogen; 2) Doseeresponse relationship and parameter value(s) for each reference pathogen; 3) Proportion of infections that result in illness (symptomatic response) (referred to hereafter as Psym) for each reference pathogen; and 4) Volume of water ingested during recreational activities. With this information, we calculated densities for each reference pathogen that, cumulatively, could have resulted in the computed swimming-associated illness rate for each beach day studied. The point estimate parameter values used for each of the variables are presented in Table 1.

Table 1 e Parameter values used in the health-based approach reverse QMRA. Reference Pathogen

Percent of Swimming-Associated Illness due to Reference Pathogen

Published Dose-Response Relationship

Rotavirus

15.6%

Beta-Poisson

Norovirus

56%

Hypergeometric

Adenovirus Cryptosporidium spp. Giardia lamblia Camplyobacter jejuni

15.6% 1.1% 7.3% 2.0%

Exponentiala Exponentialb Exponential Hypergeometric

E. coli O157:H7

0.04%

Beta-Poissonc

Salmonella enterica

0.3%

Beta-Poisson

Dose-Response Parameters

Percent of Infections Resulting in Illness

0.2531 0.4265 0.04 0.055 0.4172 0.09 0.0199 0.024 0.011 0.4 45.9 0.3126 2884

35.0% 60%d 50.0% 50.0% 45.0% 2.8%e 28% 20%

a Dose-response is for inhalation of Adenovirus 4, however the conservative assumption is employed here that all Adenoviruses are equally infectious for GI illness. b Based on analysis conducted by EPA for Long Term 2 Enhanced Surface Water Treatment Rule. c Values shown are the median values from Markov Chain Monte Carlo sample provided by Dr. Teunis. d The percent of infections resulting in illness is reported to be a function of dose. The value shown is the average of values observed during feeding study. Analysis was conducted two different ways, using both dose dependent and independent data. e The percent of infections resulting in illness is reported to be a function of dose. The value shown is the average of computed values.

4739

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

The percent of the swimming-associated illnesses attributable to each of the reference pathogens was based on data published by the CDC (Mead et al., 1999). The doseeresponse relationships were based on the peer reviewed literature (Crabtree et al., 1997; Haas et al., 1999; Regli et al., 1991; Rose and Gerba, 1991; Teunis et al., 2005, 2008a, 2008b; U.S. EPA, 2006). The percent of infections resulting in illness was based on the results of a comprehensive literature review that was conducted during the course of previous work (Soller et al., 2007a; Soller and Eisenberg, 2008) and supplemented with newer data for Norovirus (Teunis et al., 2008a). We assumed that all individuals were susceptible to Norovirus infection even though research indicates that a portion of the population may not be susceptible to infection from any particular genotype (Lindesmith et al., 2003). The volume of water ingested during recreational activities was based on data collected during pilot-scale studies conducted in swimming pools (Dufour et al., 2006). For each beach day, the probability of infection due to each reference pathogen (Pinf,rp) was assumed to be as follows: Pinf;rp ¼

Rbeach day Xrp  1000 Psymrp

[1]

where Rbeach day is the computed rate of swimming-associated illness due to all reference pathogens (expressed per 1000 swimmers), Xrp is the proportion of the swimming-associated illness rate attributable to reference pathogen rp, and Psymrp is the proportion of infections of reference pathogen rp that result in symptomatic illness. Given the above probabilities and the known mathematic forms of the doseeresponse relationships, the density of each reference pathogen for each beach day was calculated using MathCad 13 (Mathsoft Engineering and Education, Inc., 2005).

2.3.2.

POTW effluent-based QMRA approach

In the POTW effluent-based approach, pathogens were assumed to occur in the recreational waters in the same proportion as they occur in disinfected secondary effluent (secondary wastewater treatment and disinfection were in place at the various POTWs in the vicinity of the study beaches during the 2003/2004 NEEAR studies) (Fong et al., 2010; Garcia-Aljaro et al., 2004; He and Jiang, 2005; Irving

and Smith, 1981; Jimenez-Cisneros et al., 2001; Katayama et al., 2008; Koivunen et al., 2001; Lemarchand and Lebaron, 2003; Lodder and de Roda Husman, 2005; McCuin and Clancy, 2006; Rao et al., 1987; Rose et al., 2004; Soller et al., 2007b; Stampi et al., 1993). For human faecal-impacted waters this assumption is reasonable if a primary source of contamination in the waterbody was POTW effluent and the travel time was relatively short from the POTW discharge(s) to the recreational waterbodies compared to the time required for substantial inactivation of the reference pathogens. Similar to the health-based approach, concentrations of each of the reference pathogens were computed for each of the beach days. Average densities for each of the reference pathogens were then computed based on all beach days. Densities of the reference pathogens in POTW (disinfected secondary) effluent were required as input in addition to the information that was required for the health-based approach. To this end, a preliminary literature review was conducted. Based on the results of that review, point estimate parameter values were selected (Table 2). For each pathogen in the POTW effluent on each beach day, the density of the pathogen at a recreation site, Ci, was equal to the density in the POTW effluent, Yi, scaled by an attenuation factor, A: Ci ¼ AYi

[2]

with the simplifying assumption that the attenuation factor is the same for all pathogens and that the ingested dose of each pathogen is equal to A Yi V, hence, the number of illnesses per 1000 swimmers is given by: Nill =1000 ¼

m X

psym;i  fi ðAYi V; a.Þ

[3]

i¼1

where fi is the doseeresponse function for reference pathogen (rp in equation (1)) i, with parameter(s) a, ., evaluated at dose A Yi V, and psym is as defined above for each reference pathogen (Psymrp). Equation (3) was evaluated to find the attenuation parameter, A, given Nill (from epidemiology studies), Yi (effluent densities reported in the literature), the doseeresponse function for each reference pathogen and the probability of

Table 2 e Estimated densities of reference pathogens in disinfected secondary effluent. Reference Pathogen

Estimated Mean Concentration in Chlorinated Secondary Effluent

Rotavirus Norovirus

10 pfu/L 1000 qPCR genome copies/L

Adenovirus

10/L

Cryptosporidium spp. Giardia lamblia Camplyobacter jejuni E. coli O157:H7 Salmonella enterica

40 oocysts/L 12.8 cysts/L 100/L 2.5 stx2 gene carrying bacteria/L 100/L

Summary Justification

Rao et al., 1998 Katayama et al., 2008; Lodder and de Roda Husman, 2005 Irving and Smith, 1981; He and Jiang, 2005; MWRDGC, 2008; Fong et al., 2010 McCuin and Clancy, 2006 Rose et al., 2004; Soller et al., 2007 Stampi et al., 1993 Garcia-Aljaro et al., 2005 Koivunen et al., 2001; Lemarchand and Lebaron, 2003; Jimenez-Cisneros et al., 2001

4740

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

symptomatic illness given infection. The reference pathogen densities were then determined using equation (2) (code available upon request).

Incubation time and illness duration in swimmers and nonswimmers were evaluated to evaluate the plausibility of the results. Time-to-onset of illness was used as a proxy for incubation period for illnesses due to each waterborne pathogen. The time-to-onset of illness and the duration of illness were determined for individuals who reported illnesses within 10e12 days following their beach visit during the freshwater epidemiology studies. Time-to-onset of illness was defined as the time between the beach visit and date on which the occurrence of the first gastrointestinal symptom was reported. Illness duration was defined as the difference between the first and last GI symptom or the date of the telephone interview if symptoms were still occurring at the time of the interview. Since those who still had illness at the telephone interview had illness for an unknown duration beyond they were considered right-censored. Right censoring was accounted for by exponentially fitting and extending the Kaplanemeier product limit survivor curve to zero and determining the area under the curve (Selvin, 1998). Average duration was then determined by beach and contact. A total of 392 non-swimmers (no water contact), 244 waders (contact without body immersion), and 844 swimmers (contact that included body immersion) reported GI illnesses within 10e12 days of their beach visit during the freshwater epidemiology studies. GI illness rates were higher for swimmers (8.3%) than non-swimmers (6.0%) during these studies (Wade et al., 2008). GI illness in swimmers is a combination of GI illness that would occur in the absence of swimming (background) and GI illness that is attributable to water-based recreational activities. Thus, the distribution of the time-to-onset of GI illness (incubation period) in swimmers is the sum of two components, the time-to-onset that non-swimmers exhibit and a complementary portion due to the pathogens that were in the waterbodies evaluated. The time-to-onset of illness data for non-swimmers (Fig. 1) and the predicted pathogens for the epidemiology studies were used to estimate the distribution of the time-to-onset of GI illness in swimmers. That distribution was then compared to the time-to-onset of GI illness in swimmers observed during the water epidemiology studies (Fig. 1) to evaluate the feasibility of the predictions about the pathogens present during the water epidemiology studies described in the previous sections. A simulation-based approach was used to estimate the distribution of time-to-onset of GI illness for swimmers. Consistent with the outcome of the POTW-based QMRA approach (discussed below), it was assumed that swimmers were exposed to Norovirus during their recreational activities. The time-to-onset of illness from exposure to Norovirus was modeled as a triangular distribution with minimum, mode, and maximum of 0.5, 1, and 5 days, respectively, using data from Atmar et al. (2008) (Fig. 2). During each of 5000 realisations, the time-to-onset of GI illness for non-swimmers observed during the NEEAR studies

Proportion of Observations

Plausibility evaluation

0.18 0.16 Non-swimmers Swimmers

0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0

1

2

3

4

5

6

7

8

9

10

11

12

Time-to-Onset (days)

Fig. 1 e Observed time-to-onset of GI illness for nonswimmers and swimmers during NEEAR studies.

was sampled with replacement and the time-to-onset of illness for Norovirus was randomly sampled from the triangular distribution. The probability that the time-to-onset of GI illness was drawn from the background distribution, was assigned a value of 0.72 and the probability that the time-to-onset was drawn from the Norovirus distribution, a value of 0.28 (based on the ratio of GI illness rate in non-swimmers to swimmers from Wade et al., 2008).

3.

Results

3.1. Swimming-associated illness rates from the NEEAR studies The 78 daily average mean Enterococcus qPCR values were used to generate the swimming-associated GI illness rate among all swimmers. The expected value of the swimming-associated illness rates for each of the study days is provided in Table 3. The average swimming-associated GI illness rates during the NEEAR studies (defined as the difference between the proportion ill among swimmers to the proportion ill among 0.35

0.30

Proportion of Observations

2.4.

0.20

0.25

0.20

0.15

0.10

0.05

0.00 0

1

2

3

4

5

6

7

8

9

10

11

12

Time-to-Onset (days)

Fig. 2 e Time-to-onset of GI illness for Norovirus illnesses.

4741

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

non-swimmers) ranged from 26.0/1000 at Silver Beach to 36.8/ 1000 at West Beach with an overall average of 30.6/1000.

3.2. Comparison of health-based and effluent-based QMRA results The results of the Health-based QMRA approach (Table 4) indicate that: 1) a large proportion of the illnesses was attributable to a small diversity of pathogens, with Norovirus accounting for more than half of the observed illnesses; 2) Campylobacter appears to result in a relatively high proportion of infections, but a low proportion of illnesses; and 3) Salmonella densities appear to be high, but within values observed in the literature. The POTW effluent-based approach (Table 4) indicate that: 1) the majority of illnesses appear to be caused by only a few reference pathogens present in the effluent, with Norovirus by far the most important reference pathogen used; and 2) estimated pathogen densities at the beaches were within ranges reported in the literature. Both approaches indicated that relatively few reference pathogens could account for the vast majority of illnesses at the freshwater sites impacted by human contamination. For the Health-based approach the three enteric viruses and Giardia appear to be the pathogens of concern, and for the POTW effluent-based approach Norovirus represents the pathogen(s) of primary concern. For reference pathogens other than Norovirus, the POTW effluent approach yielded lower estimates of reference pathogen densities, with the difference being particularly pronounced for Salmonella. The best estimator of the combinations of reference pathogens that were present during the NEEAR studies may be bracketed by these two methods (that is, the true combination was probably a mixture of the two methods studied with contamination partially from POTW effluent and partially from other human sources). This assertion is consistent with a cause-and-effect relationship between pathogens and illnesses. If the illnesses are present, the pathogens must also be present (health approach) and if the pathogens are present, there is also the possibility of illness (POTW approach). If the reference pathogen set is missing a significant pathogen, the cause-and-effect model will be incomplete and findings might be biased. Given the data underlying both the health-based and POTW effluent approaches, we believe the reference pathogen set is adequate for characterizing health effects and water quality at the study beaches.

3.3.

Plausibility evaluation

The time-to-onset of GI illness and the duration of those illnesses for non-swimmers, waders, and swimmers from the Great Lakes epidemiology studies are summarized in Table 5. Review of those data indicates that the time-to-onset of illness for swimmers (body immersion) was slightly shorter than for non-swimmers and that this observation holds for individual beaches and across all beaches. The duration of illness appears to be similar across water contact groups and beaches. The results of the simulations estimating the time-to-onset of illness in swimmers were also compared to the observed time-to-onset of GI illness for swimmers observed during the NEEAR studies (Fig. 3). Compared to the swimmer/non-

Table 3 e Expected value of the swimming-associated illness rates for each of the NEEAR study beach days (# illnesses per 1000 swimmers). Date 1-Jun-03 7-Jun-03 8-Jun-03 14-Jun-03 15-Jun-03 21-Jun-03 22-Jun-03 28-Jun-03 29-Jun-03 4-Jul-03 5-Jul-03 6-Jul-03 12-Jul-03 13-Jul-03 19-Jul-03 20-Jul-03 26-Jul-03 27-Jul-03 2-Aug-03 3-Aug-03 9-Aug-03 10-Aug-03 16-Aug-03 23-Aug-03 24-Aug-03 30-Aug-03 31-Aug-03 6-Sep-03 7-Sep-03 13-Sep-03 14-Sep-03

Huntington Beach

West Beach

20.0 21.4 30.3 40.7 33.9 23.4 24.9 4.4 18.2 32.8 32.3 30.0 29.3 19.7 13.9 22.3 26.9 17.3 27.7 26.3 23.9 38.1 39.3 33.4

Washington Park

39.2 34.2 39.4 50.4 36.4 31.8 34.0 28.4 30.1 14.4 24.2 24.3 25.1 27.9 64.5 43.3 54.6 54.3 45.8 33.5

47.3 50.1 38.6 38.2 32.4 30.0 14.9 52.3 35.1 24.3 20.4 22.8 27.5

26-Jun-04 27-Jun-04 3-Jul-04 4-Jul-04 5-Jul-04 10-Jul-04 11-Jul-04 17-Jul-04 18-Jul-04 24-Jul-04 25-Jul-04 31-Jul-04 1-Aug-04 7-Aug-04 8-Aug-04 14-Aug-04 15-Aug-04 21-Aug-04 22-Aug-04 3-Sep-04 4-Sep-04 5-Sep-04 6-Sep-04 Average

Silver Beach

26.0

14.4 27.1 40.8 25.8 28.6 32.3 39.8 16.3 19.4 27.0 18.1 26.8 31.6 20.1 33.7 33.6 26.0 28.2 26.6 35.7 34.7 33.5 36.8

28.2

swimmer comparison (Fig. 1), these two distributions show striking similarity, supporting the QMRA estimates. Although this analysis does not provide as solid a validation as would direct comparison with observed pathogen densities during

4742

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

Table 4 e Summary of QMRA results Pathogen

Health-Based Approach

POTW Effluent-Based Approach

Illness rate/ 1000 swimmers

Infection Rate/ 1000 swimmers

Estimated mean concentration (organisms/L)

Illness Rate/ 1000 swimmers

Infection Rate/ 1000 swimmers

Estimated mean concentration (organisms/L)

30.6 4.8 17.1 4.8 0.3 2.2 0.6 0.01 0.09

Unknown 13.6 28.6 9.5 0.7 4.9 24.6 0.05 0.44

NA 0.7 2.1 0.7 0.2 7.6 1.0 0.2 122

30.6 0.3 29.7 0.3 0.2 0.01 0.1 0.001 0.0003

Unknown 0.7 49.5 0.5 0.5 0.03 3.8 0.01 0.001

NA 0.04 3.8 0.04 0.15 0.05 0.4 0.01 0.4

All Rotavirus Norovirusa Adenovirus Cryptosporidium spp. Giardia lamblia Camplyobacter jejuni E. coli O157:H7 Salmonella enterica a Genome copies per liter.

beach days (pathogen densities were not measured during the epidemiologic studies), it makes good use of available data and is one approach to evaluate the feasibility of the QMRA results.

3.4.

Sensitivity analyses

In addition to the analyses presented above (referred to as the “base analysis” hereafter), a series of sensitivity analyses were conducted on four key models parameters: the volume of water ingested during recreational activities; the densities of the reference pathogens in POTW (disinfected secondary) effluent; the Enterococcus indicator e health relationship from the NEEAR studies; and the proportion of swimming-associated illnesses attributable to each reference pathogen.

3.4.1.

Volume of water ingested

In the base analysis, the average volume ingested during recreation was a point estimate of 33 mL. We pursued a simple univariate sensitivity analysis in which the likely range of the volume of water ingested was explored (10 mLe150 mL based on best professional judgment and results from the literature). Keeping all other variables constant, the volume of water ingested was set at the lower and upper likely values, and the POTW effluent-based QMRA analysis was rerun. The results of these analyses indicated that the proportion of illnesses attributable to each of the reference pathogens was insensitive to the volume of water ingested. However, the densities of the reference pathogens at the point of exposure (the beach) and predicted attenuation values do vary linearly with the volume of water ingested.

3.4.2.

Densities of the reference pathogens in POTW effluent

In the base analysis, mid-range densities of reference pathogens in chlorinated secondary effluent were used as point estimates in the POTW effluent-based analysis. Given the results from the base analysis and the relatively low densities of pathogens predicted to be present during the beach days, we pursued a simple univariate sensitivity analysis in which lower and upper bound, literature-based densities of the reference pathogens in disinfected secondary effluent were used as input (one at a time) to determine if changes in these input values would substantially change the insights achieved through the use of point estimates. The results indicate that the mixture of pathogens that was predicted to make up the illnesses observed during the freshwater epidemiology studies was relatively consistent. The reference pathogens that were predicted to cause the highest percentage of illnesses were Norovirus, adenovirus, rotavirus and, Cryptosporidium. This is an interesting finding, given the wide range of pathogen densities reported in the literature. The relative importance of Campylobacter increased under conditions in which Norovirus densities were at the low end of those found during the preliminary literature review.

3.4.3. Proportion of swimming-associated illnesses attributable to each reference pathogen In the health-based analysis, point estimates from the U.S. Centers for Disease Control and Prevention (CDC) (Mead et al., 1999) were used to estimate the proportion of infections attributable to each reference pathogen. To estimate the total number of foodborne illnesses caused by known pathogens, CDC

Table 5 e Summary of time-to-onset and duration of illness observed in Great Lakes epidemiology studies. Water Contact Group Non-swimmer (No Contact) Wader (Contact without Body Immersion) Swimmer (Body immersion) All groups

Data Summarizede (days)

Huntington Beach

Silver Beach

West Beach

Washington Park

Total by Contact Type

Onset of Illness (Avg) Duration of illness (Avg) Onset of Illness (Avg) Duration of illness (Avg) Onset of Illness (Avg) Duration of illness (Avg) Avg Onset of Illness by Beach Avg Duration of Illness by Beach

4.6 4.3 5.4 3.2 3.7 3.0 4.5 2.4

4.6 3.5 4.8 3.4 4.2 3.2 4.4 2.5

5.7 3.4 4.2 5.1 4.1 4.2 4.4 2.8

4.5 3.8 4.1 3.0 3.6 3.2 3.9 2.6

4.7 3.9 4.7 3.7 4.0 3.5 4.3 2.5

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

in a manner parallel to the sensitivity analyses described above. We reran the base analysis in two alternative forms, corresponding to lower and upper 95% CI for the indicator e health relationship, respectively; yielding 20.0 and 41.4 illnesses per 1000 swimmers (Table 6). The results indicate that for the POTW effluent-based QMRA approach, the relative proportion of illnesses due to each of the reference pathogens remains essentially the same as that predicted by the base analysis.

0.20

Proportion of Observations

0.18 0.16 Predicted Observed

0.14

4743

0.12 0.10 0.08 0.06

4.

0.04 0.02 0.00 0

1

2

3

4

5

6

7

8

9

10

11

12

Time to Onset (days)

Fig. 3 e Observed and predicted time-to-onset of GI illness for swimmers.

determined the number of reported cases for each pathogen, adjusted to account for underreporting, and estimated the proportion of illnesses specifically attributable to foodborne transmission. The degree of underreporting is typically not available for most pathogens. For Salmonella enterica, a pathogen that typically causes non-bloody diarrhoea, the degree of underreporting was estimated at 38-fold. For E. coli O157:H7, a pathogen that typically causes bloody diarrhoea, the degree of underreporting was estimated at 20-fold from data in the literature. Using these data as reference points, CDC used an underreporting factor of 38 for pathogens that cause primarily non-bloody diarrhoea (e.g., Salmonella, Campylobacter spp.) and 20 for pathogens that cause bloody diarrhoea (e.g., E. coli O157:H7, Shigella sonnei). For pathogens that typically cause severe illness (i.e., Clostridium botulinum, Listeria monocytogenes), they used a multiplier of 2, on the assumption that most cases come to medical attention. Numerical sensitivity analyses were not conducted because data were not available to support selection of alternative values. Given the uncertainties in the CDC data, the proportion of swimming-associated illnesses attributable to each reference pathogen was also relatively uncertain, with the greatest uncertainty probably occurring for pathogens with relatively low contributions to overall non-food illness (Salmonella and Campylobacter). Nevertheless, unless the total estimated cases for each of these reference pathogens reported by CDC are extremely inaccurate, it seems reasonable to infer that the overall trends indicated by the previously summarized simulation results are likely to hold even given the substantial uncertainties that exist.

3.4.4.

Indicator e health relationship

In the base analysis, the swimming-associated GI illness rate among all swimmers was estimated as a function of daily average Enterococcus qPCR CCE based on a linear regression model which adjusted for important confounding factors. Confidence intervals (CI) about the regression line describing the relationship between Enterococcus qPCR and GI illness were previously published (Wade, 2008). We investigated the potential impact of the uncertainty associated with the indicator e health relationship

Discussion

QMRA was used to identify reference pathogens that could reasonably account for the illnesses observed during the 2003/ 2004 recreational water epidemiology studies on the Great Lakes (NEEAR studies). The sites for those epidemiology studies were selected specifically because of their proximity to treated sewage discharges. Although non-human sources undoubtedly also contributed Enterococcus to the recreational sites, human sources were assumed to dominate the risk (Schoen and Ashbolt, 2010; Soller et al., 2010). Two approaches were selected to estimate the range and mix of pathogens that could have been present during the NEEAR studies. In the Health-based approach, swimming-associated illnesses were assumed to occur in the same proportion as known illnesses occur in the US due to all non-foodborne sources. In the POTW effluent-based approach, pathogens were assumed to occur in the recreational waters in the same relative proportion as estimated in disinfected secondary effluent. Without site and temporally specific pathogen and indicator data and/or other supporting information, it is impossible to know which of the two approaches produces more accurate estimates of the pathogen densities that were present during the NEEAR epidemiology studies. However, provided that the study sites were indeed impacted by human wastewater sources, together these approaches coupled with the QMRA methodology applied here, provide a powerful context for understanding the potential public health implications of the epidemiology study results. Viruses have long been suspected as possible etiologic agents responsible for swimming-associated illness (Cabelli et al., 1982; WHO, 2003). Our results strongly support enteric viruses as being the predominant etiologic agents for freshwaters impacted by human sources. In addition to the QMRA results and the time-to-onset analyses that were consistent with Norovirus infections, water samples collected during the 2004 epidemiology studies were reported to contain human adenoviruses at both Silver Beach and Washington Park Beach (Wong et al., 2009). Interestingly, Wade et al. (2008) observed a stronger association between Enterococcus qPCR CE and GI illness than with enterococci estimated by a culturing method, and speculated that since molecular markers are less affected by sewage treatment, the PCR assay may better reflect persistent enteric pathogens such as viruses rather than pathogenic bacteria or protozoan parasites. The results of our study together with the facts that enteric viruses may be highly infectious even at low doses (Teunis et al., 2008a) and are relatively resistant to standard sewage treatment processes (Haramoto et al., 2006; Laverick et al., 2004; Lodder and de Roda Husman, 2005; Pusch et al., 2005; van den Berg et al., 2005)

4744

NA 0.05 5.3 0.05 0.21 0.07 0.53 0.013 0.5 Unknown 1.0 66.9 0.7 0.6 0.04 7.2 0.004 0.002 41.4 0.4 40.1 0.4 0.3 0.02 0.20 0.0011 0.0004 NA 0.02 2.4 0.02 0.10 0.03 0.24 0.006 0.2 Unknown 0.5 32.4 0.3 0.3 0.02 1.6 0.002 0.0009 20.0 0.2 19.5 0.2 0.1 0.01 0.05 0.0005 0.0002 NA 0.04 3.8 0.04 0.15 0.05 0.4 0.01 0.4 Unknown 0.7 49.5 0.5 0.5 0.03 3.8 0.003 0.001 a Genome copies per liter.

30.6 0.3 29.7 0.3 0.2 0.01 0.11 0.0008 0.0003 All Rotavirus Norovirusa Adenovirus Cryptosporidium spp. Giardia lamblia Camplyobacter jejuni E. coli O157:H7 Salmonella enterica

Estimated mean concentration (organisms/L) Infection Rate/1000 swimmers Estimated mean concentration (organisms/L) Infection Rate/1000 swimmers Illness Rate/1000 swimmers

Infection Rate/1000 swimmers

Estimated mean concentration (organisms/L)

Illness Rate/1000 swimmers

Lower 95% CI: Indicator Health Relationship POTW Effluent-Based Approach pathogen

Table 6 e Summary of sensitivity analysis for indicator health relationship for the POTW-based approach reverse QMRA.

Illness Rate/1000 swimmers

Upper 95% CI: Indicator Health Relationship

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

highlight the potential importance of human enteric viruses in general, and Norovirus in particular as etiologic agents of concern in recreational waters impacted by human faecal sources. The finding from a recent report from primary treated wastewater-impacted marine waters is also consistent with our findings (Silva et al., 2009). Several potential limitations of our analysis include: 1) pathogen(s) with properties very different than the reference pathogen set could be responsible for a significant proportion of illnesses observed in the epidemiological studies; 2) discrimination between adult and child infection rates (particularly for enteric viruses) was not possible; 3) it was not feasible to address potential differential response of sensitive sub-populations to exposure to the reference pathogens, 4) the potential implications of differential viability between indicators and pathogens in the freshwater environment are not accounted for, and 5) the potential importance of coinfectivity from multiple pathogens is unknown. Looking more closely at enteric viruses, most rotavirus and adenovirus GI illnesses are associated with children of <5 years of age. This fact may account for some of the differences noted between the Health- and POTW effluent-based approaches for rotavirus (illness rates of 4.8 and 0.3/1000, respectively). In this case the POTW effluent-based derivation appears to be better aligned with the current understanding of rotavirus pathology and epidemiology. Given the availability of rotavirus vaccine, rates of rotavirus infection may change in both the child and overall population. When data become available, it might be possible to account for changes in the population response to rotavirus exposure either via exclusion of rotavirus from the reference pathogen list or via change in the proportion of rotavirus infections progressing to illness. Risks from adenoviruses may warrant further consideration in the future, as the risks are highly uncertain due to extreme variability and/or uncertainty in effluent concentrations and the mismatch between the doseeresponse relationship (for adenovirus types 4 and 7 being respiratory infection via inhalation (Couch et al., 1966) whereas our interests here were for ingestion and GI illnesses). Adenoviruses 40 and 41 are the major GI infectious members, yet previous QMRA studies have also used the respiratory adenovirus doseeresponse relationship (Crabtree et al., 1997; van Heerden et al., 2005a, 2005b) due to lack of another doseeresponse relationship for adenovirus, which has a relatively high probability of infection for virion exposure. Moreover, recent work indicates that adenovirus concentrations in secondary and tertiary treated wastewater effluent may be several orders of magnitude higher than previously reported (Fong et al., 2010; He and Jiang, 2005; Irving and Smith, 1981). At the highest reported densities of adenovirus in wastewater effluent with the use of the respiratory-based doseeresponse relationship, adenoviruses appear to be important relative to the other pathogens. How these results compare to the probability that swimmers acquire GI adenovirus infections via oral ingestion is unknown. For the purposes of this work, we believe that use of the inhalation model with middle of the reported range for adenovirus density is a reasonable first approach. For the parasitic protozoa, the rates of cryptosporidiosis and giardiasis are higher in day-care aged children and their

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

caretakers than the rest of the population, with recreational swimming pool exposure being an important contributing factor (Hlavsa et al., 2005; Laupland and Church, 2005). High rates of asymptomatic infection are also common (Turabelidze et al., 2007). Therefore, depending on the demographics of the recreating population, the POTW effluent-based estimates of protozoan-infections may be more representative of the number of excess illness than the Health-based approach. Like several other reference pathogens, the number of Cryptosporidium oocysts and Giardia cysts required to cause a high probability of infection is quite low, thus we reason that the difference in parasitic protozoan infection rates among children and adults is not a driving factor in the analysis presented above.

5.

Conclusions

The principal finding from this work is that relatively few reference pathogens can account for the vast majority of the observed swimming-associated GI illnesses during the 2003/ 2004 recreational water epidemiology studies conducted on the Great Lakes. For the Health-based approach the three enteric reference viruses and Giardia appear to represent the pathogens of concern, and for the POTW effluent-based approach Norovirus alone may represent the primary concern. Finally, the results from the evaluation of the time-to-onset of illness and the sensitivity analyses strongly support these findings.

Acknowledgements The research described in this article was funded by the U.S. EPA Office of Water, Office of Science and Technology under contract #EP-C-07-036 to Clancy Environmental Consulting, Inc. This work has been subject to formal Agency review, but does not necessarily reflect the views of the Agency, and no official endorsement should be inferred. The authors gratefully acknowledge the valuable contributions of Mary Schoen for discussions leading up to this manuscript and critical review of early drafts, Quanlin Li for summarizing epidemiologic data, and Shamima Akhter and Larry Wymer for their critical review.

references

Ashbolt, N.J., Reidy, C., Haas, C.N., 1997. Microbial health risk at Sydney’s coastal bathing beaches. In: 17th Australian Water and Wastewater Association Meeting. AWWA, Melbourne, pp. 104e111. Ashbolt, N.J., Schoen, M.E., Roser, D.J., Soller, J.A., 2010. Predicting pathogen risk to aid beach management: the real value of quantitative microbial risk assessment (QMRA). Water Res. 44 (16), 4692e4703. Atmar, R.L., Opekun, A.R., Gilger, M.A., Estes, M.K., Crawford, S.E., Neill, F.H., Graham, D.Y., 2008. Norwalk virus shedding after experimental human infection. Emerg. Infect. Dis. 14 (10), 1553e1557. van den Berg, H., Lodder, W., van der Poel, W., Vennema, H., de Roda Husman, A.M., 2005. Genetic diversity of noroviruses in raw and treated sewage water. Res. Microbiol. 156 (4), 532e540.

4745

Brooks, J.P., Tanner, B.D., Gerba, C.P., Haas, C.N., Pepper, I.L., 2005. Estimation of bioaerosol risk of infection to residents adjacent to a land applied biosolids site using an empirically derived transport model. J. Appl. Microbiol. 98 (2), 397e405. Buchanan, R.L., Lammerding, A.M., Clarke, I.R., van Schothorst, M., Roberts, T.A., 1998. Potential application of risk assessment techniques to microbiological issues related to international trade in food and food products. J. Food Prot. 61 (8), 1075e1086. Buchanan, R.L., Smith, J.L., Long, W., 2000. Microbial risk assessment: dose-response relations and risk characterization. Int. J. Food Microbiol. 58 (3), 159e172. Cabelli, V.J., 1983. Health Effects Criteria for Marine Recreational Waters. USEPA, Cincinnati, OH. Cabelli, V.J., Dufour, A.P., McCabe, L.J., Levin, M.A., 1982. Swimming-associated gastroenteritis and water quality. Am. J. Epidemiol. 115, 606e616. CDC, 2006. Surveillance for waterborne-disease outbreaks associated with recreational water e United States, 2003e2004. MMWR Surveill. Summ. 55 (SS-12), 1e24. Couch, R.B., Cate, T.R., Douglas Jr., R.G., Gerone, P.J., Knight, V., 1966. Effect of route of inoculation on experimental respiratory viral disease in volunteers and evidence for airborne transmission. Bacteriol. Rev. 30 (3), 517e529. Crabtree, K.D., Gerba, C.P., Rose, J.B., Haas, C.N., 1997. Waterborne adenovirus: a risk assessment. Water Sci. Technol. 35 (11e12), 1e6. Dudley, R.H., Hekimian, K.K., Mechalas, B.J., 1976. A scientific basis for determining recreational water quality criteria. J. Water Pollut. Control Fed. 48 (12), 2661e2677. Dufour, A.P., 1984. Health Effects Criteria for Fresh Recreational Waters. USEPA, Cincinnati, OH. Dufour, A.P., Evans, O., Behymer, T., Cantu, R., 2006. Water ingestion during swimming activities in a pool: a pilot study. J. Water Health 4 (4), 425e430. Eisenberg, J.N.S., Soller, J.A., Scott, J., Eisenberg, D.M., Colford, J.M., 2004. A dynamic model to assess microbial health risks associated with beneficial uses of biosolids. Risk Anal. 24 (1), 221e236. Eisenberg, J.N.S., Moore, K., Soller, J.A., Eisenberg, D., Colford Jr., J.M., 2008. Microbial risk assessment framework for exposure to amended sludge projects. Environ. Health Perspect. 116 (6), 727e733. Farber, J.M., Ross, W.H., Harwig, J., 1996. Health risk assessment of Listeria monocytogenes in Canada. Int. J. Food Microbiol. 30 (1e2), 145e156. Fleisher, J.M., Kay, D., Salmon, R.L., Jones, F., Wyer, M.D., Godfree, A.F., 1996. Marine waters contaminated with domestic sewage: nonenteric illnesses associated with bather exposure in the United Kingdom. Am. J. Public Health 86 (9), 1228e1234. Fong, T.T., Phanikumar, M.S., Xagoraraki, I., Rose, J.B., 2010. Quantitative detection of human adenoviruses in wastewater and combined sewer overflows influencing a Michigan river. Appl. Environ. Microbiol. 76 (3), 715e723. Fuhs, O.W., 1975. A probabilistic model of bathing beach safety. Sci. Total Environ. 4 (2), 165e175. Garcia-Aljaro, C., Muniesa, M., Jofre, J., Blanch, A.R., 2004. Prevalence of the stx2 gene in coliform populations from aquatic environments. Appl. Environ. Microbiol. 70 (6), 3535e3540. Gerba, C.P., 2000. Assessment of enteric pathogen shedding by bathers during recreational activity and its impact on water quality. Quant. Microbiol. 2, 55e68. Gerba, C.P., Rose, J.B., Haas, C.N., Crabtree, K.D., 1996. Waterborne rotavirus: a risk assessment. Water Res. 30 (12), 2929e2940. Haas, C.N., 1983. Effect of effluent disinfection on risks of viral disease transmission via recreational water exposure. J. WPCF 55 (8), 1111e1116.

4746

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

Haas, C.N., Rose, J.B., Gerba, C.P., 1999. Quantitative Microbial Risk Assessment. John Wiley & Sons Inc. Haile, R.W., Witte, J.S., Gold, M., Cressey, R., McGee, C., Millikan, R.C., Glasser, A., Harawa, N., Ervin, C., Harmon, P., Harper, J., Dermand, J., Alamillo, J., Barrett, K., Nides, M., Wang, G.Y., 1999. The health effects of swimming in ocean water contaminated by storm drain runoff. Epidemiology 10 (4), 355e363. Haramoto, E., Katayama, H., Oguma, K., Yamashita, H., Tajima, A., Nakajima, H., Ohgaki, S., 2006. Seasonal profiles of human noroviruses and indicator bacteria in a wastewater treatment plant in Tokyo, Japan. Water Sci. Technol. 54 (11e12), 301e308. He, J.W., Jiang, S., 2005. Quantification of enterococci and human adenoviruses in environmental samples by real-time PCR. Appl. Environ. Microbiol. 71 (5), 2250e2255. van Heerden, J., Ehlers, M.M., Grabow, W.O.K., 2005a. Detection and risk assessment of adenoviruses in swimming pool water. J. Appl. Microbiol. 99, 1256e1264. van Heerden, J., Ehlers, M.M., Vivier, J.C., Grabow, W.O.K., 2005b. Risk assessment of adenoviruses detected in treated drinking water and recreational water. J. Appl. Microbiol. 99, 926e933. Hlavsa, M.C., Watson, J.C., Beach, M.J., 2005. Cryptosporidiosis surveillance e United States 1999e2002 and giardiasis surveillance e United States, 1998e2002. MMWR Surveill. Summ. 54 (1), 1. ILSI, 1996. A conceptual framework to assess the risks of human disease following exposure to pathogens. Risk Anal. 16 (6), 841e848. ILSI, 2000. Revised Framework for Microbial Risk Assessment. ILSI Press, Washington, DC. Irving, L.G., Smith, F.A., 1981. One-year survey of enteroviruses, adenoviruses, and reoviruses isolated from effluent at an activated-sludge purification plant. Appl. Environ. Microbiol., 51e59. Jimenez-Cisneros, B.E., Maya-Rendon, C., Salgado-Velazquez, G., 2001. The elimination of helminth ova, faecal coliforms, Salmonella and protozoan cysts by various physicochemical processes in wastewater and sludge. Water Sci. Technol. 43 (12), 179e182. Jones, F., Kay, D., Stanwell-Smith, R., Wyer, M., 1991. Results of the first pilot-scale controlled cohort epidemiological investigation into the possible health effects of bathing in Seawater at Langland Bay, Swansea. Water Environ. J. 5 (1), 91e98. Katayama, H., Haramoto, E., Oguma, K., Yamashita, H., Tajima, A., Nakajima, H., Ohgaki, S., 2008. One-year monthly quantitative survey of noroviruses, enteroviruses, and adenoviruses in wastewater collected from six plants in Japan. Water Res. 42 (6e7), 1441e1448. Kay, D., Fleisher, J.M., Salmon, R.L., Jones, F., Wyer, M.D., Godfree, A.F., Zelenauch-Jacquotte, Z., Shore, R., 1994. Predicting likelihood of gastroenteritis from sea bathing: results from randomized exposure. Lancet 344 (8727), 905e909. Koivunen, J., Lanki, E., Rajala, R.L., Siitonen, A., HeinonenTanski, H., 2001. Determination of Salmonellae from municipal wastewaters. Wat. Sci. Tech. 43 (12), 221e224. Laupland, K.B., Church, D.L., 2005. Population-based laboratory surveillance for Giardia sp. and Cryptosporidium sp. infections in a large Canadian health region. BMC Infect. Dis. 5, 72. Laverick, M.A., Wyn-Jones, A.P., Carter, M.J., 2004. Quantitative RTPCR for the enumeration of noroviruses (Norwalk-like viruses) in water and sewage. Lett. Appl. Microbiol. 39 (2), 127e136. Lemarchand, K., Lebaron, P., 2003. Occurrence of Salmonella spp. and Cryptosporidium spp. in a French coastal watershed: relationship with fecal indicators. FEMS Microbiol. Lett. 218 (1), 203e209. Lindesmith, L., Moe, C., Marionneau, S., Ruvoen, N., Jiang, X., Lindbland, L., Stewart, P., LePendu, J., Baric, R., 2003. Human

susceptibility and resistance to Norwalk virus infection. Nature Med. 9 (5), 548e553. Lodder, W.J., de Roda Husman, A.M., 2005. Presence of noroviruses and other enteric viruses in sewage and surface waters in The Netherlands. Appl. Environ. Microbiol. 71 (3), 1453e1461. McCuin, R.M., Clancy, J.L., 2006. Occurrence of Cryptosporidium oocysts in US wastewaters. J. Water Health 4 (4), 437e452. Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Shapiro, C., Griffin, P.M., Tauxe, R.V., 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5 (5), 607. Mena, K., Gerba, C.P., Haas, C.N., Rose, J.B., 2003. Risk assessment of waterborne coxsackievirus. J. Am. Water Works Assoc. 95 (7), 122e131. Pru¨ss, A., 1998. Review of epidemiological studies on health effects from exposure to recreational water. Int. J. Epidemiol. 27 (1), 1e9. Pusch, D., Oh, D.Y., Wolf, S., Dumke, R., Schroter-Bobsin, U., Hohne, M., Roske, I., Schreier, E., 2005. Detection of enteric viruses and bacterial indicators in German environmental waters. Arch. Virol. 150 (5), 929e947. Rao, V.C., Metcalf, T.G., Melnick, J.L., 1987. Removal of indigenous rotaviruses during primary settling and activated-sludge treatment of raw sewage. Water Res. 21 (2), 171e177. Regli, S., Rose, J.B., Haas, C.N., Gerba, C.P., 1991. Modeling the risk from Giardia and viruses in drinking-water. J. Am. Water Works Assoc. 83 (11), 76e84. Riley, S., Fraser, C., Donnelly, C.A., Ghani, A.C., Abu-Raddad, L.J., Hedley, A.J., Leung, G.M., Ho, L.M., Lam, T.H., Thach, T.Q., Chau, P., Chan, K.P., Lo, S.V., Leung, P.Y., Tsang, T., Ho, W., Lee, K.H., Lau, E.M., Ferguson, N.M., Anderson, R.M., 2003. Transmission dynamics of the etiological agent of SARS in Hong Kong: impact of public health interventions. Science 300 (5627), 1961e1966. Rose, J.B., Mullinax, R.L., Singh, S.N., Yates, M.V., Gerba, C.P., 1987. Occurrence of rotaviruses and enteroviruses in recreational waters of Oak Creek, Arizona. Water Res. 21 (11), 1375e1381. Rose, J.B., Haas, C.N., Regli, S., 1991. Risk assessment and control of waterborne giardiasis. Am. J. Public Health 81 (6), 709e713. Rose, J.B., Nowlin, H., Farrah, S.R., Harwood, V., Levine, A., Lukasik, J., Menendez, P., Scott, T.M., 2004. Reduction of Pathogens, Indicator Bacteria, and Alternative Indicators by Wastewater Treatment and Reclamation Processes. Water Environment Research Foundation Report 00-PUM-2T. Rose, J.B., Gerba, C.P., 1991. Assessing potential health risks from viruses and parasites in reclaimed water in Arizona and Florida, USA. Water Sci. Technol. 23 (10/2), 2091e2098. Roser, D.J., Ashbolt, N.J., Davies, C.M., Glamore, W.C., Hawker, K.M., Miller, B.M., 2007. Application of TMDL and risk assessment principles for pathogen management at an urban recreation lake. In: Watershed management to meet water quality standards and TMDLS, 4th Conference Proceedings, San Antonio, Texas, The American Society of Agricultural and Biological Engineers (ASABE): 420e426. Schets, F.M., Engels, G.B., Evers, E.G., 2004. Cryptosporidium and Giardia in swimming pools in the Netherlands. J. Water Health 2 (3), 191e200. Schoen, M.E., Ashbolt, N.J., 2010. Assessing pathogen risk to swimmers at non-sewage impacted recreational beaches. Environ. Sci. Technol. 44 (7), 2286e2291. Selvin, S., 1998. Modern Applied Biostatistical Methods Using SPlus. Oxford University Press, New York, pp. 432e433. Silva, A.M., Vieira, H., Martins, N., Granja, A.T., Vale, M.J., Vale, F.F., 2009. Viral and bacterial contamination in recreational waters: a case study in the Lisbon bay area. J. Appl. Microbiol. Soller, J.A., Eisenberg, J.N.S., 2008. An evaluation of parsimony for microbial risk assessment models. Environmetrics 19 (1), 61e78.

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 7 3 6 e4 7 4 7

Soller, J.A., Olivieri, A.W., Crook, J., Cooper, R.C., Tchobanoglous, G., Parkin, R.T., Spear, R.C., Eisenberg, J.N., 2003. Risk-based approach to evaluate the public health benefit of additional wastewater treatment. Environ. Sci. Technol. 37 (9), 1882. Soller, J.A., Eisenberg, J., DeGeorge, J., Cooper, R., Tchobanoglous, G., Olivieri, A., 2006. A public health evaluation of recreational water impairment. J. Water Health 4 (1), 1e19. Soller, J.A., Seto, E., Olivieri, A.W., 2007a. Microbial Risk Assessment Interface Tool, Final Project Report, Water Environment Research Foundation Project 04-HHE-3. Soller, J.A., Seto, E.Y., Olivieri, A.W., 2007b. Application of Microbial Risk Assessment Techniques to Estimate Risk due to Exposure to Reclaimed Waters, WateReuse Foundation Report WRF-04-011. Soller, J.A., Schoen, M.E., Bartrand, T., Ravenscroft, J., Ashbolt, N.J., 2010. Estimated human health risks from exposure to recreational waters impacted by human and non-human sources of faecal contamination. Water Res. 44 (16), 4674e4691. Stampi, S., Varoli, O., Zanetti, F., De Luca, G., 1993. Arcobacter cryaerophilus and thermophilic campylobacters in a sewage treatment plant in Italy: two secondary treatments compared. Epidemiol. Infect. 110 (3), 633e639. Teunis, P.F., Medema, G.J., Kruidenier, L., Havelaar, A.H., 1997. Assessment of the risk of infection by Cryptosporidium or Giardia in drinking water from a surface water source. Water Res. 31 (6), 1333e1346. Teunis, P.F., Van den Brandhof, W., Nauta, M., Wagenaar, J., Van den Kerkhof, H., Van Pelt, W., 2005. A reconsideration of the Campylobacter dose-response relation. Epidemiol. Infect. 133 (4), 583e592. Teunis, P.F., Moe, C.L., Liu, P., Millar, S.E., Lindesmith, L., Baric, R.S., Le Pendu, J., Calderon, R.L., 2008a. Norwalk virus: how infectious is it? J. Med. Virol. 80 (8), 1468e1476. Teunis, P.F., Ogden, I.D., Strachan, N.J., 2008b. Hierarchical dose response of E. coli O157:H7 from human outbreaks incorporating heterogeneity in exposure. Epidemiol. Infect. 136, 761e770.

4747

Turabelidze, G., Lin, M., Weiser, T., Zhu, B.P., 2007. Communitywide outbreak of cryptosporidiosis in rural Missouri associated with attendance at child care centers. Arch. Pediatr. Adolesc. Med. 161 (9), 878e883. U.S. EPA, 1986. Ambient Water Quality Criteria for Bacteria. EPA440/5-84-002. Office of Water, Washington, DC. U.S. EPA, 2006. National Primary Drinking Water Regulations: Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR); Final Rule. U.S. EPA, 2007. Critical Path Science Plan for the Development of New or Revised Recreational Water Quality Criteria. Wade, T.J., Pai, N., Eisenberg, J., Colford, J.M., 2003. Do US EPA water quality guidelines for recreational waters prevent gastrointestinal illness? A systematic review and metaanalysis. Environ. Health Perspect. 111 (8), 1102e1109. Wade, T.J., Calderon, R.L., Sams, E., Beach, M., Brenner, K.P., Williams, A.H., Dufour, A.P., 2006. Rapidly measured indicators of recreational water quality are predictive of swimming-associated gastrointestinal illness. Environ. Health Perspect. 114 (1), 24e28. Wade, T.J., Calderon, R.L., Brenner, K.P., Sams, E., Beach, M., Haugland, R., Wymer, L., Dufour, A.P., 2008. High sensitivity of children to swimming-associated gastrointestinal illness: results using a rapid assay of recreational water quality. Epidemiology 19 (3), 375e383. WHO, 2003. Guidelines for Safe Recreational-water Environments, Coastal and Fresh-waters, vol. 1. World Health Organization, Geneva. WHO, 2004. Guidelines for Drinking-water Quality, third ed., vol. 1. World Health Organization, Geneva. Wong, M., Kumar, L., Jenkins, T.M., Xagoraraki, I., Phanikumar, M. S., Rose, J.B., 2009. Evaluation of public health risks at recreational beaches in Lake Michigan via detection of enteric viruses and a human-specific bacteriological marker. Water Res. 43 (4), 1137e1149. Zmirou, D., Pena, L., Ledrans, M., Letertre, A., 2003. Risks associated with the microbiological quality of bodies of fresh and marine water used for recreational purposes: summary estimates based on published epidemiological studies. Arch. Environ. Health 58 (11), 703e711.