Eighty Wistar rats were randomized into two groups. In group 1 vesicosigmoidostomy with proximal colostomy was performed, in group 2, vesicosigmoidostomy. The total tumor incidence did not significantly (group 1 10/40, 25%; group 2 13/40, 32.5%). The tumor spectrum differed, with more adenocarcinomas in group 2 (11/40, 27.5% vs 4/40, 10%; P = 0.047) and urothelial carcinomas only in group 1 (5/40, 2.5%). One hundred and ten other Wistar rats were randomized into three groups. Animals in group A received vesicoileosigmoidostomy, group B, two-step vesicosigmoidostomy with initial separation of urine and the urocolonic anastomosis, group C, vesicosigmoidostomy. Significantly fewer adenocarcinomas were observed in group A (2/40, 5%) than in group B (16/40, 40%; P < 0.002) and group C (9/30, 30%; P < 0.007). These results indicate a similar cancer risk in all continent forms of urinary diversion, at least via colon. Ileal interposition seems to be an effective carcinoma prophylaxis following ureterosigmoidostomy. The proliferative instability at the urointestinal anastomosis is crucial for the pathogenesis and prophylaxis of this form of carcinogenesis, whereas urine seems to play only a minor role.