UC Davis

Human pathogens have been observed in municipal drinking water systems. Microbial intrusions due to distribution system deficiencies are a recognized cause of acute gastrointestinal illness (AGI) outbreaks. Yet, their importance as a cause of sporadic and endemic AGI is poorly understood. The WAHTER (Water And Health Trial for Enteric Risk) study sought to characterize health impacts associated with virus intrusions into drinking water distribution systems, as well as with virus contamination of source groundwater.

The WAHTER study was designed as a cohort epidemiological trial with cross-over intervention, taking place in 14 small Wisconsin communities relying on untreated groundwater. Six human enteric virus types were enumerated by RT-qPCR in the source groundwater and household tap water of the study communities during four three-month periods in 2006-2007. AGI incidence in the communities was reported through weekly health questionnaires. Each community received intervention in the form of UV disinfection at active wellheads for one of the study years. During UV intervention viral contamination in source groundwater was virtually eliminated, so that viruses detected at taps could be ascribed to intrusions into distribution systems.

Within the context of the WAHTER study, this dissertation focuses on: the validation of a method to concentrate viruses from water using glass wool filtration (Chapter 1); a method to estimate virus numbers in environmental samples using integrated cell culture-qPCR (Chapter 2); the association between distribution system maintenance or failure events and viral contamination in distribution systems (Chapter 3); and estimating non-epidemic AGI incidence associated with virus intrusions into distribution systems (Chapter 4).

The association between virus levels at household taps and distribution system events was tested using generalized linear regression. The occurrence of pipe installations was significantly associated with increased virus levels in distribution systems, while hydrant flushing was significantly associated with reduced virus loads (Chapter 3).

AGI associated with distribution systems was estimated using a Monte Carlo-based static risk assessment framework. Results suggest that virus intrusions into drinking water distribution systems can be a significant contributor to non-epidemic AGI. Distribution system risk was highly variable, reflecting the sporadic nature of contamination events (Chapter 4).