Surgical repair of the anal sphincters after previous trauma is generally successful. In earlier publications, a protective colostomy was recommended but in most recent series colostomy is omitted. We have been through both phases and this is the first comparative study done on 82 consecutive repairs: 45 with colostomy from 1977 to 1986 (Group I) and 37 without colostomy from 1986 to 1992 (Group II). Causes of trauma were obstetrical: 50, surgical: 24 and violence: 5. Apart from colostomy related morbidity, postoperative complication rates were similar in the two groups. Results were graded excellent, good, fair or poor according to continence to solids, to liquids and soiling. Good and excellent results were obtained in 82% (Group I) and 87% (Group II) after a mean follow-up duration of 42 and 23 months respectively. Furthermore there was no difference between Group I and II in the rate of good/excellent results for cases who had undergone prior repairs (98% v. 100%) and also when the duration of incontinence was more than 10 years (71% v. 83%). We conclude that colostomy is not a determinant factor in the outcome and is therefore not required, avoiding all colostomy related morbidity and disability.