A total of 95 patients with stage Ta/T1 superficial bladder cancer was eligible for intravesical bacillus Calmette-Guerin (BCG) immunoprophylaxis according to the standard high risk criteria for tumor recurrence and progression. Of these patients 50 agreed to undergo treatment while the remaining 45 refused any intravesical therapy and served as concurrent nonrandomized controls. The patient and tumor characteristics in the 2 groups of patients were identical. All patients underwent complete transurethral resection of the bladder tumor. Therapy consisted of 120 mg. weekly instillations of BCG (Danish 1331 strain) for 6 weeks. Among the 50 patients 33 received BCG initially while 17 received it after failure of intravesical chemotherapy. Responses were evaluated according to standard criteria. Recurrence was noted in 24 patients (48%) in the BCG group compared to 38 (84.4%) in the control group. The recurrence rates per year and per 100 patient-months were significantly reduced in the BCG arm compared to the control arm. The mean interval to first recurrence and the mean recurrence interval were significantly increased in the BCG arm compared to the control arm. The relative risk of recurrence in the BCG group was 0.62 versus 1.63 in the control group. Subgroup analysis showed significant benefit of BCG for patients with single as well as multiple, stages Ta and T1, and grades II and III tumors. Comparison with pretreatment controls in the BCG group revealed a significant reduction in the recurrence rate in those patients after treatment with BCG, which was not seen in the control group. The benefit of BCG was seen in those who received BCG initially as well as in those who received it after failure of intravesical chemotherapy. Multivariate analysis of prognostic factors showed that this benefit in the BCG group was related only to the treatment, while other prognostic parameters, such as tumor stage, grade, number and so forth were not independent prognostic variables. The rate of progression to muscle invasion was not significantly different in the 2 treatment groups. The relapse-free survival in the BCG group was 35.4% at 60 months compared to 11.2% in the control group (p < 0.001). The side effects of BCG therapy were mild, brief and easily controlled with conservative measures.