Click on the links below to access all the ArticlePlus for this article.
Please note that ArticlePlus files may launch a viewer application outside of your web browser.
It would be useful to monitor recreational waters continuously for human pathogens as a way to prevent swimming-associated infections. However, there is considerable difficulty and expense associated with testing for the vast number of potentially pathogenic microorganisms. Instead, fecal indicator bacteria such as Escherichia coli or Enterococcus are used to assess the microbial safety of recreational waters. These indicator bacteria are generally not harmful but can be a marker for the presence of sewage and human feces, and the health risks that result from such exposures. Because currently used methods require 24 to 48 hours to obtain results, monitoring does not immediately detect changes in exposure, leading to delays in notifying beach-goers of possible risks.
A faster method of measuring water quality could improve protection of public health by reducing the time between exposure measurement and management decisions, potentially providing same-day results before most beach-goers enter the water. We previously reported that a faster method of measuring fecal indicator bacteria using quantitative polymerase chain reaction (QPCR) showed promise in its ability to predict swimming-associated gastrointestinal (GI) illness.1 After sample collection and transport, the QPCR method can be performed in 3 hours or less. With further improvements this may be shortened to 2 hours or less.2
We expand on our previous analysis to include 2 additional freshwater beaches, an assessment of nonenteric illnesses (upper respiratory illness [URI], rash, eye irritations, and earaches), and separate analyses by age.
We conducted a prospective study of visitors to freshwater Great Lake beaches on Lake Michigan and Lake Erie during the summers of 2003 and 2004. The data collection methods have been described previously.1 In brief, we attempted to enroll all beach-goers between 11:00 am and 5:00 pm during summer weekends and holidays. We excluded unaccompanied minors (below 18 years) or those who could not speak English or Spanish. We interviewed volunteers as they were leaving the beach to ascertain information about swimming and other activities. Ten to 12 days later, one of the adults in the household was interviewed by telephone about health symptoms experienced by participating household members. All subjects provided oral consent. The study procedures were approved by the Institutional Review Board for the Centers for Disease Control and Prevention.
Human-derived pollution sources generally cause the most health concern,3 and beaches with such pollution were the focus of these studies. In 2003, we conducted studies at West Beach (on Lake Michigan in Indiana Dunes National Seashore in Indiana) and Huntington Beach (on Lake Erie near Cleveland, OH). In 2004, we studied 2 additional Lake Michigan Beaches: Silver Beach, near St. Joseph, Michigan, and Washington Park Beach in Michigan City, Indiana. The range of fecal indicator-bacteria concentration at these beaches is related to contamination by effluent from sewage treatment plants. Water quality at each beach was influenced by point-source tributaries that received combined treated sewage treatment discharges from communities with populations of at least 38,000 and with flow rates of over 10 million gallons per day (see Appendix A, available with the online version of this article, for additional details). These sewage plants provided secondary treatment as well as disinfection with chlorine or ultraviolet radiation during the summer.
Water Sample Collection and Sample Analysis
Water samples were tested for fecal indicator bacteria Enterococcus and Bacteroides using QPCR. Because of problems in the sensitivity of the Bacteroides QPCR assay, the data in 2004 were insufficient to assess this indicator in relation to health effects. Samples were also tested for Enterococcus using EPA Method 1600,4 one of the culture-based methods currently recommended by the EPA for recreational freshwater monitoring.5
We collected water samples at 8:00 am, 11:00 am, and 3:00 pm, along 3 transects perpendicular to the shoreline—one sample in waist-high water (1 m deep) and one in shin-high water (0.3 m deep). Transects were located at least 60 m apart to encompass the swimming area. Because rock jetties at Huntington Beach prevented free circulation of water, we collected 4 additional samples at each sampling time to better characterize the water quality. Following collection, samples were placed in coolers and maintained on ice at 1 to 4°C. Analyses of samples by Method 1600 for Enterococcus were performed by local laboratories within 6 hours of collection. Samples were filtered for QPCR analysis within 6 hours of collection. To ensure consistency across the 4 beaches, the filters were frozen and sent on dry ice by overnight express for analysis by EMSL Analytical, Inc. Laboratory (Westmont, NJ).
The QPCR method used in this study has been previously described.1,6 In brief, organisms in water samples were collected by membrane filters, total DNA was extracted, and polymerase chain reaction (PCR) amplification of a genus-specific DNA sequence of Enterococcus was carried out using the TaqMan PCR product detection system. The reactions were performed in a thermal cycling instrument (SmartCycler System, Cepheid, Sunnyvale, CA) that automated the detection and quantitative measurement of the fluorescent signals produced by probe degradation during each cycle of amplification. Ratios of the target sequences in a test sample were compared with a calibrator sample using an arithmetic formula, referred to as the Comparative Cycle Threshold Method.7 These ratios were converted to measurements of calibrator cell equivalents in test samples through the use of calibrator samples processed in the same manner as the test samples and containing a known quantity of the target organism cells. Results are reported in QPCR cell equivalents (QPCR CE) per 100 mL of original sample.
At each sampling time we recorded environmental conditions, including air and water temperature, cloud cover, rainfall, wind speed and direction, wave height, beach population density, boats, animals (number and type), and debris.
We assessed 5 endpoints, defined a priori, and similar to those previously studied.8–13
1. “Gastrointestinal illness” (GI illness) was defined as any of the following: diarrhea (three or more loose stools in a 24-hour period); vomiting; nausea and stomach ache; nausea or stomach ache, and interference with regular activities (missed time from work or school, or missed other regular activities as a result of the illness).
2. “Upper respiratory illness” (URI) was defined as any 2 of the following: sore throat, cough, runny nose, cold, or fever.
3. “Rash” was defined as a rash or itchy skin.
4. “Eye ailments” were defined as either eye infection or watery eye.
5. “Earache” was defined as earache, ear infection, or runny ears.
Consistent with some other studies,8,10,11 URI and GI illness were not restricted to persons with fever, since infections can produce these illnesses without fever (eg, E. coli 0157:H7 and norovirus infections). We were also concerned about the accuracy of self-reported low-grade fever.
People who were ill within 3 days before their beach visit were excluded for the outcome with which they had been afflicted. We examined various definitions of GI illness including diarrhea (three or more loose stools in a 24-hour period) alone and GI illness with complications (defined as missing regular activities, using medications, or visiting a health provider as a result of a GI symptom).
Definition of Swimming
“Swimmers” were those who reported immersing their body to their waist or higher. In our previous analysis, immersion to the waist showed a pattern of risk similar to head immersion.1 Nonswimmers were defined as those who reported no contact with water. Those entering the water not up to their waist were classified as “waders.”
Because QPCR CE were highly skewed, raw data were log-transformed (base 10). The arithmetic mean of the log-transformed values was used to summarize water quality at a given day, time, or location. We previously used a maximum-likelihood method to impute results for samples below the limit of detection.1 However, QPCR CE from 2004 beaches were not as well approximated by a log-normal distribution, making imputation based on this exact distributional assumption questionable. Furthermore, we had concern that partial inhibition may have been responsible for some nondetected results, making the imputed detection limits incorrect. We therefore excluded samples below the limit of detection from the calculation of averages. However, this choice of method for dealing with the limit of detection did not affect the results (see Appendix B, available with the online version of this article). We focused the analyses primarily on 2 summary measures: the daily average of all samples and the average of the 8:00 am samples. The daily average represented average water quality at a beach on a particular day. The 8:00 am average was used to determine if morning water quality was predictive of illness among swimmers exposed later that day—an important consideration for assessing the utility of a faster method such as QPCR for water quality evaluation. Analyses using depth-specific averages were also conducted and results are shown in Appendix B. Analysis of variance models were used to explore the relationship between log10 QPCR CE with beach, collection date, time, and sample depth.
To account for correlated environmental measurements as potential confounders of the swimming and health effects relationships, we used principal-components analysis to produce summary components. The 5 principal components that accounted for the majority of the variability (54%) were included in health effects regression models. These 5 components were beach-goer density, temperature (water and air), rainfall, wind direction and debris, and wind speed and wave height. To avoid data loss when one or more of the environmental observations were missing (18 of 85 days), principal components were imputed using best-subset regression.14
We used generalized linear-regression models to evaluate the association between water quality and health effects. Logistic regression models were used to describe the strength of the association between the QPCR CE measures and incidence of illness among swimmers. Models using an identity link and a binomial error structure (linear model) were used to directly estimate the attributable risk15 (swimmer risk minus nonswimmer risk), which we refer to as “swimming-associated illness.” Although the linear and logistic models produced similar results, the linear models allowed direct estimation of the attributable risk, which is often considered a more meaningful and direct statement of risk.5 Nonswimmers were included in models and were assigned water quality exposures of zero. Indicators for “swimming” and “beach” were included in all models. Log-linear models were used to estimate the adjusted cumulative incidence ratio associated with swimming (without regard to water quality).15 Robust estimates of variance were used to account for the nonindependence of observations within household.16–18
Covariates strongly associated with swimming, water quality or illness, or those considered by investigators to be potential confounding factors were considered for inclusion in regression models. These factors included age, sex, race, contact with animals, contact with other persons with diarrhea, number of other visits to the beach, any other chronic illnesses (GI, skin, asthma), digging in sand, and the first 5 principal components of the environmental/meteorological factors (described above). An indicator was also created for a festival that took place at Silver Beach, drawing 17,000 visitors to an area adjacent to the beach. For URI, rash, and eye outcomes, use of insect repellent and sun block were also considered. For each analysis, the set of covariates was reduced through a change-in-estimate procedure.19 A criterion of a 5% change was used, although this was occasionally relaxed to obtain a parsimonious model. The selection procedure generally reduced the numbers of covariates to 7 or fewer.
To evaluate heterogeneity in the indicator/illness relationship across the beaches, we graphically examined the relationship at each beach and conducted likelihood ratio tests. These tests compared models with interaction terms between beach and water quality (which allowed slopes to differ across beaches) with restricted models constrained to a single slope across the 4 beaches.
We conducted separate analysis for the age categories 0 to 10 years, 11 to 54 years, and 55 years and older. The age groups were selected a priori based on sample size and investigators' judgment.
A total of 21,015 interviews from 10,093 household groups were completed (Appendix B). Respondents at the 4 beaches differed by age, race, miles traveled to the beach and proportion of swimmers (Appendix B). Respondents were 85% white and 56% female, with a median age of 27 years. Swimmers were younger than nonswimmers (median age 19 and 35 years, respectively) but were equally likely to report rash, sore throat, vomiting, and eye irritations in the 3 days prior to the beach visit, chronic respiratory illness (eg, asthma), and chronic skin problems (Table 1). Slightly fewer swimmers compared with nonswimmers reported chronic GI conditions (2% vs. 3%), GI symptoms (other than vomiting) in the 3 days prior to the beach visit (2% vs. 3%), chronic allergies (18% vs. 21%), and consumption of red or raw meat prior to or immediately after the beach visit (8% vs. 10%). There were more female beach-goers than male in all 3 water-use groups, with the largest discrepancy among the waders (63% vs. 37%) and the smallest among the swimmers (52% vs. 48%). Most were white. The percentages of other races were similar across water-use groups, except that the percentage of Hispanic/Latino respondents was highest among the swimmers. More swimmers than nonswimmers reported using sunblock (61% vs. 40%), insect repellant (3% vs. 2%), and having had contact with animals (79% vs. 75%).
Enterococcus QPCR CE differed by beach (Table 2) and sample depth. Median QPCR CEs at shin depth were higher than waist depth (93 and 65 QPCR CE/100 mL, respectively). Collection time was not an important factor in the variability of QPCR CE, although QPCR CE levels measured at 3:00 pm were slightly higher than at 8:00 am (median QPCR CE/100 mL was 74, 78, and 80 at 8:00 am, 11:00 am, and 3:00 pm, respectively).
Twenty-five of 78 days (32%) exceeded the current geometric mean guideline value of 33 colony forming units (CFU)/100 mL Enterococcus measured by Method 1600.5 Twenty-two percent (333 of 1482) of individual samples exceeded the single sample maximum of 61 CFU/100 mL.5
The incidence of new GI illness was 7.3% (1497 of 20,414) during the 10 to 12 day follow-up period. GI illness incidence was highest among children younger than 5 years (9.0%) and lowest among those aged 55 and older (4.9%). The adjusted risk of GI illness was 1.44 times higher in swimmers than nonswimmers (95% CI = 1.27–1.64; Table 3). The risks among children aged 10 and younger, and children and adults aged 11 to 54, were similar to the pooled risk. Among those aged 55 and older, swimmers reported 2.3 times as many illnesses as nonswimmers of the same age (1.33–3.99; Table 4). Children aged 5 and younger showed a similar pattern of risk as those aged 10 years and younger, but with the exception of GI illness (1.67 [CI = 1.03–2.69]), small sample sizes prohibited making conclusions about this age group.
Approximately 5.7% of respondents reported URI. Incidence was highest in children younger than 5 (10.6%) and lowest in those aged 55 and older (2.5%). The crude incidence of URI was higher among swimmers than nonswimmers, but after adjustment there was little difference in risk (1.06 [0.90–1.24]; Table 3]. Age was a strong confounder because young respondents were both more likely both to swim and report URI. Among children aged 10 years and younger, URI risk was not elevated among swimmers (0.95 [0.66–1.38]; Table 4).
Approximately 2.7% of all respondents reported rash, with the highest incidence in children younger than 5 years (4.1%), and the lowest in those aged 55 and older (2.1%). Swimmers reported more rash than nonswimmers (1.38 [CI = 1.12–1.72]; Table 3). Rashes occurred more frequently on the upper and lower back (26%) of swimmers reporting rash than of nonswimmers reporting rash (12%).
The incidence of eye irritations and infections was 2.9%; these were reported with equal frequency by swimmers and nonswimmers (1.00 [0.81–1.24]; Table 3).
Relationships Between Water Quality and Health
The incidence of GI illness was consistently associated with Enterococcus QPCR CE exposure (Table 5, Figs. 1 and 2). Among all subjects, a 1 log10 increase in the daily QPCR CE average resulted in a 1.26 increase in the risk (odds) of GI illness (95% CI = 1.06–1.51). The relationship was stronger among children, with a similar association for those aged 10 years and younger (1.69 [1.24–2.30]), 5 and younger (1.67 [1.08–2.57]), and 2 and younger (1.65 [0.81–3.36]). The association between the 8:00 am Enterococcus QPCR CE average and GI illness was nearly identical to that of the daily average. As illustrated in Figure 1, 1000 swimmers exposed to 100 Enterococcus QPCR CE would experience an average of 34 more episodes of GI illness than nonswimmers. One thousand swimming children aged 10 and younger would experience an average of 49 more episodes than nonswimming children (Fig. 2). The associations between Enterococcus QPCR CE and GI illness were positive at each of the 4 beaches, and tests for heterogeneity indicated no difference in these relationships across the 4 beaches among all subjects (P = 0.84), or among children aged 10 and younger (P = 0.65). Crude rates of GI illness and numbers exposed are presented in Appendix C (available with the online version of this article).
As time spent in the water increased beyond 1.5 hours, the association between Enterococcus QPCR CE and GI illness also increased. Among subjects exposed at least 2 hours, the risk of GI illness associated with Enterococcus QPCR CE exposure increased (1.89 [1.07–3.35]). Children aged 5 to 10 years spent the most time in the water, an average of 1.5 hours compared with 1.2 hours for those younger than 5 years, 1.2 hours for those aged 11 to 20, and less than an hour for those older than 20.
Other illnesses did not show strong or consistent associations with Enterococcus QPCR CE. For example, rash was positively associated overall with Enterococcus QPCR CE exposure among all subjects (aOR = 1.21 [CI = 0.92–1.58]; Table 5) and particularly among children (1.58 [0.90–2.76]; Table 5). However, there was significant (P = 0.02) variation in the association across the 4 beaches, with strong positive associations at 2 of the beaches (at Silver Beach, aOR =1.78 [CI = 1.08–2.94] and at Washington Park Beach aOR = 1.60 [0.59–4.31]. At the other 2 beaches there was no evidence of an association (Huntington Beach 0.87 [0.27–2.85], and West Beach 0.92 [0.57–1.48]). The data were too sparse to reliably assess heterogeneity on the association of beach-specific rash with Enterococcus QPCR CE among children.
Enterococcus Measured by Method 1600 and GI Illness
Swimmers exposed above the guideline value of 33 CFU/100 mL had higher risks than nonswimmers or swimmers exposed below this value (Table 6). As with QPCR CE, the risks associated with Enterococcus CFU exposure were more pronounced among children aged 10 and younger.
Enterococcus QPCR CE levels were a stronger predictor of GI illness than the CFU measure. Among all subjects, a quartile increase in Enterococcus QPCR CE was associated with a 1.44 (95% CI = 1.10–1.90) increase in the odds of illness for all subjects, whereas a quartile increase Enterococcus CFU was associated with only a 1.04 (0.90–1.21) increase. Among children, a quartile increase in Enterococcus QPCR CE was associated with a 2.27 increase in the odds of illness (1.34–1.68) compared with a 1.21 (95% CI = 0.94–1.55) increase for a quartile increase in Enterococcus CFU. In models including both Enterococcus measurements, a quartile increase in the daily QPCR average and illness was strengthened (1.56 [1.14–2.12]), while the relationship between a quartile increase in CFU and GI illness was weakened (0.95 [0.79–1.13]).
A molecular method for rapid measurement of water quality (Enterococcus QPCR CE) was consistently associated with swimming-associated GI illness at 4 freshwater beaches. Furthermore, the Enterococcus QPCR CE showed that children up to age 10 years were especially susceptible to GI illness following swimming exposure. While a sensitivity among children to illness following recreational water exposure has often been hypothesized,3,20–23 this is the first study to demonstrate this sensitivity as a function of microbial water quality. At least one previous study has observed higher rates of swimming-associated illness among children, but the authors did not attribute the increased illnesses to measures of water quality.24 Children may be more likely to swallow water,25 transfer water to their mouth after exposure, or, as we observed, spend a longer time in water, resulting in a greater likelihood of contact with pathogens. Children are at increased susceptibility to infection and illness caused by several enteric pathogens.26,27 Such susceptibility may be due to differences in immune system function, hygiene, and other physiological and behavioral differences.27
We saw no evidence of increased susceptibility among those aged 55 and older, but our ability to make valid conclusions among this group was limited because they swam infrequently and reported the lowest incidence of illness. Swimmers in this age group did have a higher overall risk for GI illness compared with nonswimmers, but the relative risk may have been skewed by the low incidence of GI illness among nonswimmers.
Some of the health endpoints were nonspecific, and may have been affected by recall bias. Broad endpoints accounted for the diverse range of symptoms potentially associated with recreational water exposure but such broad symptoms may obscure more specific effects of water quality and swimming exposure. The association between Enterococcus QPCR CE and GI illness, however, was robust to different definitions (diarrhea, GI illness with complications). While swimmers may have been more likely to recall illness than nonswimmers, it is unlikely such a recall bias would occur among swimmers at varying levels of water quality. As with our previous analysis, a more stringent definition of swimming with head immersion did not substantially alter the results (data not shown).
Numerous studies have considered associations between fecal indicator bacteria and symptoms of illness. The majority of these studies have observed some association with GI illness.22,28,29 Associations between fecal indicator bacteria and nongastrointestinal (nonenteric) health conditions appear to be less consistent. Several studies9–11,13 observed associations with respiratory illness, although not all.8,12,30–33 Similar inconsistencies have been observed for skin, ear, and eye ailments.8–10,12,30,31,34 Earaches and ear infections are often associated with swimming and water exposure, but associations with specific indicator organisms have been inconsistent.8–10,30,35
Enterococcus QPCR CE was more strongly associated with illness than the currently recommended culture-based method of measuring Enterococcus. The QPCR measure may be a truer representation of fecal contamination, because it measures all Enterococcus associated with feces, not just viable cells. The molecular measurement of Enterococcus DNA provides a stable, conservative means of quantifying the level of fecal contamination, which is not subject to die-off but may mirror the dilution and dispersion of fecal material. Studies have demonstrated that pathogenic microorganisms (especially viruses and certain protists) are capable of surviving the sewage treatment process. Levels of such pathogens in treated effluent are often poorly correlated with indicator bacteria measured by cultural methods.36 Whereas fecal indicator bacteria are often nondetectable by culture methods following sewage treatment, these same bacteria can be detected by QPCR.37 A recent study found human adenoviruses at both Silver Beach and Washington Park Beach, with municipal discharges as the likely source.38
The water quality at the beaches we studied was influenced by human sources of pollution. We do not know if the relationships we observed between Enterococcus QPCR CE and GI illness can be extended to marine beaches, or to recreational waters affected by different sources of fecal contamination. Our failure to observe consistent associations between nonenteric illness and fecal indicator bacteria suggests a continuing need to investigate the causes of excess nonenteric illnesses commonly observed among swimmers.
We gratefully acknowledge the work and cooperation of the following: Karen Della Torre, Kurt Patrizi, Robert Clickner, Richard Whitman, Justin Telech, Joel Hansel, Ann Williams, Mark Murphy, Westat, Inc., the NEEAR field team, National Park Service-Indiana Dunes National Lakeshore, USGS-Lake Michigan Ecological Research Station, Cleveland Metroparks-Bradley Woods/Huntington Reservation, Cuyahoga County Board of Health, Berrien County Parks and Recreation Commission, Michigan City Parks and Recreation, La Porte County Health Department, and the U.S. Army Corps of Engineers. A special acknowledgment is reserved for all of the families and individuals who took the time to participate in the NEEAR Water Studies.
1.Wade TJ, Calderon RL, Sams E, et al. Rapidly measured indicators of recreational water quality are predictive of swimming-associated gastrointestinal illness. Environ Health Perspect
2.Siefring S, Atikovic MVE, Wymer L, et al. Improved real-time PCR assays for the detection of fecal indicator bacteria in surface waters with different instrument and reagent systems. J Water Health
3.Pond K. Water Recreation and Disease Plausibility of Associated Infections: Acute Effects, Sequelae and Mortality. London: IWA Publishing; 2005.
4.USEPA. Method 1600: Membrane Filter Test for Enterococci in Water
. Washington DC: Office of Water; 1997.
5.USEPA. Ambient Water Quality for Bacteria-1986
. Cincinnati, OH: Office of Water Regulations and Standards; 1986, p. 18.
6.Haugland RA, Siefring SC, Wymer LJ, et al. Comparison of Enterococcus
measurements in freshwater at two recreational beaches by quantitative polymerase chain reaction and membrane filter culture analysis. Water Res
7.Applied Biosystems. User Bulletin #2: ABI PRISM 7700 Sequence Detection System. Foster City, CA: Applied Biosystems Corporation; 1997.
8.Colford JM Jr, Wade TJ, Schiff KC, et al. Water quality indicators and the risk of illness at beaches with nonpoint sources of fecal contamination. Epidemiology
9.Fleisher JM, Kay D, Salmon RL, et al. Marine waters contaminated with domestic sewage: nonenteric illnesses associated with bather exposure in the United Kingdom. Am J Public Health
10.Haile RW, Witte JS, Gold M, et al. The health effects of swimming in ocean water contaminated by storm drain runoff. Epidemiology
11.McBride GB, Salmondo CE, Bandaranayake DR, et al. Health effects of marine bathing in New Zealand. Int J Environ Health Res
12.Prieto MD, Lopez B, Juanes JA, et al. Recreation risks in coastal waters: health risks associated with bathing in sea water. J Epidemiol Commun Health
13.von Schirnding YE, Kfir R, Cabelli V, et al. Morbidity among bathers exposed to polluted seawater. A prospective epidemiological study. S Afr Med J
14.Little R, Rubin D. Statistical Analysis with Missing Data
. 2nd ed. New York: Wiley; 2002.
15.Jewell NP. Statistics for Epidemiology. Texts in Statistical Science
. Vol. 58. Boca Raton: Chapman & Hall/CRC; 2004.
16.Huber P. The Behavior of Maximum Likelihood Estimates Under Non-Standard Conditions
. Berkeley, CA: University of California Press; 1967, pp. 221–233.
17.Rogers W. Regression standard errors in clustered samples. Stata Tech Bull
18.White H. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica
19.Greenland S, Rothman KJ. Modern epidemiology. Philadelphia, PA: Lippincott-Raven, 1998.
20.Alexander LM, Heaven A, Tennant A, et al. Symptomatology of children in contact with sea water contaminated with sewage. J Epidemiol Commun Health
21.Dorfman M. Testing the Waters: A Guide to Water Quality at Vacation Beaches
. Natural Resources Defense Council; 2005.
22.Wade TJ, Pai N, Eisenberg JN, et al. Do U.S. Environmental Protection Agency Water Quality Guidelines for Recreational Waters prevent gastrointestinal illness? A systematic review and meta-analysis. Environ Health Perspect
23.World Health Organization. Bathing Water Quality and Human Health: Faecal Pollution. Outcome of an Expert Consultation
. Farnham, UK: World Health Organization; 2001, p. 58.
24.Cheung WH, Chang KC, Hung RP, et al. Health effects of beach water pollution in Hong Kong. Epidemiol Infect
25.Dufour AP, Evans O, Behymer TD, et al. Water ingestion during swimming activities in a pool: a pilot study. J Water Health
26.Koehler KM, Lasky T, Fein SB, et al. Population-based incidence of infection with selected bacterial enteric pathogens in children younger than five years of age, 1996–1998. Pediatr Infect Dis J
27.Nwachuku N, Gerba CP. Health risks of enteric viral infections in children. Rev Environ Contam Toxicol
28.Pruss A. Review of epidemiological studies on health effects from exposure to recreational water. Int J Epidemiol
29.Zmirou D, Pena L, Ledrans M, et al. 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
30.Ferley JP, Zmirou D, Balducci F, et al. Epidemiological significance of microbiological pollution criteria for river recreational waters. Int J Epidemiol
31.Kueh CSW, Tam TY, Lee T, et al. Epidemiological study of swimming-associated illnesses relating to bathing-beach water quality. Water Sci Technol
32.Lee JV, Dawson SR, Ward S, et al. Bacteriophages are a better indicator of illness rates than bacteria amongst users of a white water course fed by a lowland river. Water Sci Technol
33.Seyfried PL, Tobin RS, Brown NE, et al. A prospective study of swimming-related illness. II. Morbidity and the microbiological quality of water. Am J Public Health
34.Marino F, Morinigo M, Martinez-Manzanares E, et al. Microbiological–epidemiological study of selected marine beaches in Malaga (Spain). Water Sci Technol
35.Corbett SJ, Rubin GL, Curry GK, et al. The health effects of swimming at Sydney beaches. The Sydney Beach Users Study Advisory Group. Am J Public Health
36.Harwood VJ, Levine AD, Scott TM, et al. Validity of the indicator organism paradigm for pathogen reduction in reclaimed water and public health protection. Appl Environ Microbiol
37.He JW, Jiang S. Quantification of enterococci and human adenoviruses in environmental samples by real-time PCR. Appl Environ Microbiol
38.Xagoraraki I, Kuo DH-W, Wong K, et al. Occurrence of human adenoviruses in two great lake recreational beaches. Appl Environ Microbiol
This article has been cited
Supplemental Digital Content
© 2008 Lippincott Williams & Wilkins, Inc.