Escherichia coli counts in household drinking water may or may not reliably indicate the presence of diarrheogenic pathogens originating in feces. The extent to which a bacterial indicator like E. coli predicts risks from all classes of pathogens (viruses and parasites as well as bacteria), especially in tropical waters, is uncertain. To investigate the association between E. coli in household drinking water and diarrheal diseases in Cambodia, we conducted a 22 week cohort study in a rural village in Kandal Province. Episodes of diarrhea (all) and bloody diarrhea (dysentery), water quality, water sources, and other covariates were monitored biweekly in 180 households. Households used a variety of water treatment, storage, and handling practices.
Results suggest a weak but positive association between E. coli counts in household drinking water and diarrhea and for diarrhea with blood (dysentery), after adjusting for clustering within households and within individuals over time. Compared to households with <1 E. coli/100 ml in drinking water, there was no observed increased risk for having 1–10 E. coli/100 ml (LPR = 0.98, 95% CI 0.81–1.2 for diarrheal disease; LPR = 0.75, 95% CI 0.36–1.6 for dysentery). Households with measured E. coli of 11–100/100 ml did report increased diarrhea (LPR = 1.2, 95% CI 1.1–1.3 for diarrheal disease; LPR = 1.4, 95% CI 1.0–1.8 for dysentery), as did those with 101–1,000 E. coli/100 ml (LPR = 1.2, 95% CI 1.2–1.3 for diarrheal disease; LPR = 1.2, 95% CI 1.0–1.4 for dysentery) and those with >1,000 E. coli per 100 ml sample (LPR = 1.2, 95% CI 1.1–1.2 for diarrheal disease; LPR = 1.2, 95% CI 1.0–1.3 for dysentery). Unlike the results of some previous studies, diarrheal disease risks did not increase progressively in magnitude with increasing concentration of E. coli in drinking water.