OBJECTIVE This is a study of the long-term course of surgically treated Crohn's disease designed to identify prognostic factors predictive of the time course and probability of surgical recurrence. METHODS The study is based on the records of 101 patients admitted to our institution for surgical treatment of Crohn's disease from January 1, 1970 to December 31, 1985. Follow-up was complete in 97 (96 percent) and incomplete in 4 patients. Median follow-up from the date of first operation was 13.25 years. The cumulative probability of requiring surgical treatment for recurrent disease was calculated using the life table method and further analyzed with the log-rank test and Cox regression. RESULTS The time to reoperation in this series was not significantly influenced by sex, age at onset of symptoms, age at diagnosis, age at first operation, anatomic location, and number of sites involved at the time of first operation. The only variable that had a statistically significant effect on the time to reoperation was characterization of disease at the time of operation as being perforating (P) opposed to nonperforating (NP). Median interval between the first and second intestinal operation was 1.7 years for the P group and 13 years for the NP group (P value, 0.005), and the median time between any two operations undergone during the study period was 2 years for the P group and 9.9 years for the NP group (P = 0.0002). The risk of having to undergo reoperation for recurrence was greatest during the first two years after an operation, and this was mainly because of a short time to surgical recurrence in the P group of indications. Therefore, the yearly hazard of requiring further surgery was maintained at approximately 5 percent. CONCLUSIONS The cumulative probability of requiring a reoperation for patients undergoing surgery for the P type of Crohn's disease is significantly different from that of patients with NP indications. The risk of having to undergo further surgery is particularly high during the first two years following an operation for perforating disease. The concept of a relatively aggressive perforating type of Crohn's disease and a more indolent nonperforating type is confirmed by the results of this study.