A randomized controlled trial assessing infectious disease risks from bathing in fresh recreational waters in relation to the concentration of Escherichia coli, intestinal enterococci, Clostridium perfringens, and somatic coliphages

Abstract

We performed epidemiologic studies at public freshwater bathing sites in Germany to provide a better scientific basis for the definition of recreational water quality standards. A total of 2,196 participants were recruited from the local population and randomized into bathers and nonbathers. Bathers were exposed for 10 min and had to immerse their head at least three times. Water samples for microbiological analysis were collected at 20-min intervals. Unbiased concentration-response effects with no-observed-adverse-effect levels (NOAELs) were demonstrated for three different definitions of gastroenteritis and four fecal indicator organisms. Relative risks for bathing in waters with levels above NOAELs compared with nonbathing ranged from 1.8 (95% CI, 1.2-2.6) to 4.6 (95% CI, 2.1-10.1), depending on the definition of gastroenteritis. The effect of swallowing water provided additional evidence for true dose-response relationships. Based on the NOAELs, the following guide values for water quality are suggested: 100 Escherichia coli, 25 intestinal enterococci, 10 somatic coliphages, or 10 Clostridium perfringens per 100 mL. Recreational water quality standards are intended to protect the health of those consumers who are not already immune or resistant to pathogens that may be associated with indicator organisms. In contrast to current World Health Organization recommendations, we concluded that standards should be based on rates of compliance with NOAELs rather than on attributable risks determined above NOAELs, because these risks depend mainly on the unpredictable susceptibility of the cohorts. Although in theory there is no threshold in real concentration-response relationships, we demonstrated that a NOAEL approach would be a more robust and practical solution to the complex problem of setting standards.

Figure 1

Participant flow through each stage of the trial in CONSORT format (Moher et al. 2001) for the analysis of gastroenteritis (definition GE_UK-wf) and the concentration of E. coli and intestinal enterococci. ^aDetails are explained in “Materials and Methods.” ^bNumber depends on the disease definition. ^cNumber depends on the microbiological parameter. ^dNumber depends on the disease definition and the microbiological parameter.

Figure 2

Incidence rates of gastroenteritis (definition GE_UK-wf) within 1 week after exposure to fresh recreational water in quartile and quintile categories of E. coli concentrations (A) and incidence rates below and above the presumed NOAEL of 78 E. coli /100 ml (B), depending on the accidental swallowing of water. *p < 0.05 in a chi-square test or Fisher’s exact test comparing the incidence rate of gastroenteritis in one of the specified exposure categories with the incidence rate of gastroenteritis in the group of unexposed participants (nonbathers). **p < 0.001.

Figure 3

Calculation of the disease burden from exposure to log-normal distributed intestinal enterococci concentrations (A) and the influence of cohort susceptibility and disease definition on the acceptable level of fecal water pollution if standards are based on predefined attributable risks (in this case, 3 and 5%) above an assumed identical NOAEL of 25 intestinal enterococci/100 mL (B). Details on the calculation of the PDF, the disease burden, and the aRs are explained in “Discussion.”