OBJECTIVE: To determine the prevalence and risk factors for nosocomial infections (NIs) in four Swiss university hospitals. DESIGN AND SETTING: A 1-week period-prevalence survey conducted in May 1996 in medical, surgical, and intensive-care wards of four Swiss university hospitals (900-1,500 beds). Centers for Disease Control and Prevention definitions were used, except that asymptomatic bacteriuria was not categorized as NI. Study variables included patient demographics, primary diagnosis, comorbidities, exposure to medical and surgical risk factors, and use of antimicrobials. Risk factors for NIs were determined using logistic regression with adjustment for length of hospital stay, study center, device use, and patients' comorbidities. RESULTS: 176 NI were recorded in 156 of 1,349 screened patients (11.6%; interhospital range, 9.8%-13.5%). The most frequent NI was surgical-site infection (53; 30%), followed by urinary tract infection (39; 22%), lower respiratory tract infection (27; 15%), and bloodstream infection (23; 13%). Prevalence of NI was higher in critical-care units (25%) than in medical (9%) and surgical wards (12%). Overall, 65% of NIs were culture-proven; the leading pathogens were Enterobacteriaceae (44; 28%), Staphylococcus aureus (20; 13%), Pseudomonas aeruginosa (17; 11%), and Candida species (16; 10%). Independent risk factors for NI were central venous catheter (CVC) use (odds ratio [OR], 3.35; 95% confidence interval [CI95], 2.91-3.80), admission to intensive care (OR, 1.75; CI95, 1.30-2.21), emergency admission (OR, 1.57; CI95, 1.15-2.00), impaired functional status (Karnofsky index 1-4: OR, 2.56; CI95, 1.953.17), and McCabe classification of ultimately fatal (OR, 2.50; CI95, 2.04-2.96) or rapidly fatal (OR, 2.25; CI95, 1.52-2.98) underlying condition. CONCLUSIONS: According to the results of this survey, NIs are frequent in Swiss university hospitals. This investigation confirms the importance of CVCs as a major risk factor for NI. Patient comorbidities must be taken into account to adjust for case mix in any study comparing interhospital or intrahospital infection rates.