BACKGROUND: Because of the concern for infection risk, use of sterile water has been recommended in the water bottle for endoscopic equipment, although studies evaluating prevalence of contamination of the water bottle with clinical outcomes have not been performed. METHODS: Over a 12-week period in three endoscopy rooms at a university teaching hospital, the water bottles were filled on a weekly schedule with either sterile (one room) or tap water. The water bottles were sterilized on a weekly basis with an automated endoscope washer. At the end of each week, an aliquot of the remaining water was transferred to a sterile container, and quantitative cultures for aerobic and facultative anaerobic bacteria were performed by use of a 0.01 ml calibrated loop according to standard protocols. Cultures were performed in a blinded fashion without knowledge of the water source. Follow-up was performed on all patients within 2 weeks of the procedure to determine any potential infectious complications. RESULTS: During the study period, 437 procedures were performed (203 endoscopy, 68 colonoscopy, 38 sigmoidoscopy, 128 endoscopic retrograde cholangiopancreatography). Of a total of 36 cultures (12 sterile), the results of nine (25%) were positive, including three bottles where sterile water was used. Bacterial isolates included five Flavobacterium sp., four Acinetobacter sp., two Pseudomonas sp., and one Stenotrophomonas maltophilia. Colony counts ranged from 900 to more than 10,000 per ml. On follow-up no patient had development of a clinical infection from any of these organisms. CONCLUSIONS: Bacterial growth in the water bottle was infrequent, consisted predominantly of nonpathogenic organisms, and was not associated with clinical complications. Our pilot study suggests that the use of tap water as compared with sterile water may be practical as well as provide cost savings.