In a prospective study including 137 consecutive catheterised patients in a medical intensive care unit, the following variables were analysed as possible risk factors for catheter-associated bacteriuria, defined as a quantitative culture with > or = 10(5) organisms/ml: age, sex, simplified acute and physiologic score at admission, duration of catheterisation, diabetes mellitus, immunosuppression, neurologic disorders and prior systemic antibiotic exposure during hospitalisation. The frequency of catheter-associated bacteriuria was 30.7%. By multivariate analysis, female sex (odds ratio [OR], 5.1; 95% confidence interval [CI], 1.9-13.5; P=0.001) and a duration of catheterisation >11 days (OR, 19.4; 95% CI, 5.5-68.7; P=0.0001) were risk factors for catheter-associated bacteriuria, and prior antibiotic exposure was a protective factor (OR, 0.06; 95% CI, 0.019-0.21; P=0.0001).