Epidemiological significance of microbiological pollution criteria for river recreational waters.

A retrospective follow-up study was conducted during the summer of 1986 in the French Ardèche basin in order to assess the relationship between swimming-related morbidity and the bacteriological quality of the recreational water. 5737 tourists in eight holiday camps were questioned as to the occurrence of illness and their bathing habits during the week preceding the interviews. The rate-ratio contrasting swimmers and non-swimmers for total morbidity is 2.1 (1.8-2.4) = 95% confidence interval); gastrointestinal illness is the major type of morbidity and differs significantly between the two groups (RR = 2.4 (1.9-3.0) for total gastrointestinal cases; RR = 2.3 (1.7-3.2) for 'objective' gastrointestinal cases). Faecal streptococci (FS) are best correlated to gastrointestinal morbidity, using direct linear regression models. Faecal coliforms (FC) are not as good predictors of the risk. The concentration of faecal streptococci above which the 'objective' gastrointestinal morbidity among bathers is significantly greater than among non-bathers is 20 FS/100 ml. Swimmers suffer skin ailments much more frequently than non-swimmers (RR = 3.7 (2.4-5.7]; although the relationship may be artefactual, this type of morbidity is well correlated with the concentration of faecal coliforms, aeromonas and pseudomonas. This study provides epidemiological data on which to base microbiological standards for river recreational waters dependent on what might be considered as an 'acceptable' risk.