The management of penetrating colon injury has been frequently debated in the literature, yet few reports have evaluated primary closure versus diverting colostomy in similarly injured patients. Diverting colostomy is the standard of care when mucosal penetration is present, but primary closure in civilian practice has generally had excellent results, although it has been restricted to less severely injured patients. Because the degree of injury may influence choice of treatment in modern practice, various indices of injury severity have been proposed for assessment of patients with penetrating colon trauma. As yet, however, there has been no cross-comparison of repair type versus injury severity. A retrospective study 76 patients who sustained penetrating colon trauma between January 1, 1979 and December 31, 1985 and who survived for at least 24 hours was conducted. Different preferences among attending surgeons and a more aggressive approach to the use of primary closure during the years of study led to an essentially random use of primary closure and diverting colostomy for moderate levels of colon injury, with mandatory colostomy reserved for the most serious injuries. Primary closure was performed in 37 patients (three having resection and anastomosis), and colostomy was performed in 39 patients. Severity of injury was evaluated by the Injury Severity Score (ISS), Penetrating Abdominal Trauma Index (PATI), and the Flint Colon Injury Score. Complications and outcome were evaluated as a function of severity of injury, and primary closure and colostomy were compared. Demographic profiles of the two groups did not differ regarding age, sex, mechanism of injury, shock, or delay between injury and operation. The mortality rate was 2.6% for each group. Major morbidity, including septic complications, occurred in 11% of the patients of the primary closure group and in 49% of those of the colostomy group. When PATI was less than 25, the Flint score was less than or equal to 2, or when the ISS was less than 25, primary closure resulted in fewer complications than did colostomy. Of the injury severity indices examined, the PATI most reliably predicted complications and specifically identified patients who whose outcome would be good with primary repair. These results suggest that the use of primary closure should be expanded in civilian penetrating colon trauma and that, even with moderate degrees of colon injury, primary closure provides an outcome equivalent to that provided by colostomy. In addition, the predictive value of the PATI suggests that it should be included along with other injury severity indices in trauma data bases.