Over a 10-year period we reviewed the records of 370 consecutive patients with potentially penetrating abdominal wounds (48 gunshot and 322 stab wounds). Selective non-operative management for abdominal stab wounds was introduced, guided by serial physical examination. In our study the terms therapeutic and non-therapeutic laparotomies were used, the latter comprising negative as well as insignificant findings. Initially diagnostic peritoneal lavage and local wound exploration were used, but these methods were later abandoned. Mortality was 10.2% and morbidity 25% in the group of gunshot wounds and 1.2% and 8.6% in the group of stab wounds. During this period the rate of laparotomies for stab wounds decreased from 55% to 30%, while the rate of non-therapeutic laparotomies decreased from 24% to 0% in the last year. Delayed laparotomy did not cause death or increase morbidity. Our results support the concept of selective non-operative management of abdominal stab wounds using repeated physical examination. Peritoneal perforation and haemoperitoneum should not be an indication for routine laparotomy.