To compare the effectiveness of bacterial interference versus placebo in preventing urinary tract infection (UTI). The main outcome measure was the numbers of episodes of UTI/patient-year. Randomization was computer generated, with allocation concealment by visibly indistinguishable products distributed from a core facility. The healthcare providers and those assessing the outcomes were unaware of the group allocation. Adult patients (n = 65) with neurogenic bladder after spinal cord injury and a history of recurrent UTI were randomized in a 3:1 ratio to receive either Escherichia coli HU2117 or sterile saline. Urine cultures were obtained weekly during the first month and then monthly for 1 year. The patients were evaluable if they remained colonized with E. coli HU2117 for >4 weeks (experimental group). The trial is closed to follow-up. Of the 59 patients who received bladder inoculations, 27 were evaluable (17 in the experimental group and 10 in the placebo group). The 2 study groups had comparable clinical characteristics. Of 17 patients colonized with E. coli HU2117 and the 10 control patients, 5 (29%, 95% confidence interval 0.11-0.56) and 7 (70%, 95% confidence interval 0.35-0.92) developed >1 episode of UTI (P = .049; 1-sided Fisher's exact test), respectively. The average number of episodes of UTI/patient-year was also lower (P = .02, Wilcoxon rank sum test) in the experimental (0.50) than in the control group (1.68). E. coli HU2117 did not cause symptomatic UTI. Bladder colonization with E. coli HU2117 safely reduces the risk of symptomatic UTI in patients with spinal cord injury. Effective, but less complex, methods for achieving bladder colonization with E. coli HU2117 are under investigation. Published by Elsevier Inc.