In a prospective study, 131 of 1458 patients acquired 136 urinary-tract infections (defined as greater than 10(5) colony-forming units per milliliter) during 1474 indwelling bladder catheterizations. Seventy-six patients (25 infected and 51 noninfected) died during hospitalization; death rates were 19 per cent in infected patients and 4 per cent in noninfected patients. Multiple logistic regression analysis demonstrated that seven of 21 prospectively monitored variables were associated with mortality among the catheterized patients. The adjusted odds ratio for mortality between those who acquired infection and those who did not was 2.8 (95 per cent confidence limits, 1.5 to 5.1). The acquisition of infection as not associated with the severity of underlying disease; among patients who died, infections occurred in 38 per cent of those classified as having nonfatal underlying disease (15 of 39) and in 27 per cent of those classified as having fatal disease (10 of 37). Twelve deaths may have been caused by acquired urinary-tract infections. Two patients had urinary-tract pathogens in premortem blood cultures. Another 10 died with clinical pictures compatible with serious infection, but no diagnostic cultures were performed. We conclude that the acquisition of urinary-tract infection during indwelling bladder catheterization is associated with nearly a threefold increase in mortality among hospitalized patients, but the reason for this association is not yet clear.