Between 1970 and 1988, 1379 patients with Crohn's disease were treated at the University of Chicago. Of these, 639 (mean age, 32.5 years; 322 men, 317 women) required at least one surgical procedure. The most common indications for operation were failure of medical treatment (n = 215, 33%), presence of a fistula (n = 154, 24%), and bowel obstruction (n = 141, 22%). A fistula was the most common intraoperative Crohn's-related complication. In 582 patients (92%), a resection was necessary, with primary anastomosis in 416 (65%), a temporary stoma in 124 (20%), and a permanent stoma in 42 (7%). The remaining 57 patients underwent diverse procedures (stricturoplasty, bypass, and so on). Two patients (0.3%) died. Follow-up data was obtained in 95%. One hundred eighteen patients developed recurrence requiring reoperation. The recurrence rate was 20% at 5 years and 34% at 10 years. The recurrence involved a permanent stoma or a previous anastomosis in 62 patients (afferent limb in 46, efferent in 16). In the 391 patients without previous surgery for Crohn's disease, a covariate analysis was performed to determine those variables significantly associated with recurrence. Variables included demographic data, findings at operation, surgical procedures, and histopathologic characteristics. The analysis revealed that the number of sites involved was the only variable that was significantly associated with the intra-abdominal recurrence rate (p less than 0.001). The annualized risk of recurrence was 1.6% for patients with single-site involvement and 4% for those with multiple-site involvement. Perineal disease was associated with a significantly higher risk of local recurrence than any other site (p less than 0.02). A subanalysis of 236 patients with single-site involvement but no previous operation allowed us to study the influence of site on indications for surgery and type of operative procedure. Failure of medical treatment was the most common indication for all sites. In contrast the site involved influenced the procedure: resection and primary anastomosis was feasible in 88% of jejunoileal and terminal ileal cases and a temporary ileostomy was necessary in only 12%. No patients with small bowel localization required a permanent stoma. A resection with primary anastomosis was feasible in only 32% of patients with colonic disease. The remaining two thirds of patients required either a temporary or a permanent stoma. It is concluded that multisite involvement is associated with 2.5 times the rate of recurrence of single-site disease, while the presence of perineal disease has a significantly higher incidence of local recurrence.(ABSTRACT TRUNCATED AT 400 WORDS)