We assessed the efficacy of periodic instillations of hydrogen peroxide into urinary drainage systems in the prevention of catheter-associated bacteriuria in a prospective and randomized clinical study of 668 patients with indwelling urethral catheters. Bacteriuria was documented in 68 (10%) of the 668 patients after a mean duration of four days of catheterization. There was no difference between the hydrogen peroxide group and the control group in the mean duration of catheterization before the onset of bacteriuria, in the attack rate for bacteriuria, or in the spectrum of etiologic agents recovered. Bag contamination with the same organism responsible for bacteriuria preceded infection in only five (7%) of the 68 patients, three patients using hydrogen peroxide and two in the control group. We conclude that infections arising intraluminally from contamination of the drainage bag are uncommon among catheterized patients and that the periodic instillation of disinfectants into closed sterile drainage systems is not effective in reducing the incidence of catheter-associated bacteriuria.