BACKGROUND: Despite the well documented morbidity associated with its reversal, Hartmann's procedure remains the favoured option in patients with complicated diverticular disease in the presence of diffuse peritonitis. A prospective study was conducted to determine whether primary anastomosis with diverting colostomy constitutes a valid alternative to the Hartmann procedure. METHODS: Between 1994 and 1998, all patients with diffuse peritonitis due to perforated diverticulitis of sigmoid origin underwent resection and primary anastomosis with diverting colostomy. Restoration of colonic continuity was programmed six weeks later, after verification of the anastomose by gastrografin enema. The group included 5 men and 15 women with a mean age of 72 years (32-97 years). The ASA classification of the patients was as follows: ASA II (n = 2), ASA III (n = 12), ASA IV (n = 3), ASA V (n = 3). The mean delay between onset of symptoms and surgery was 74 hours (8-215 hours). RESULTS: Operative mortality and morbidity was 15% (n = 3) and 50% respectively. No patients showed signs of suture disruption and this was confirmed by routine radiological controls of the anastomoses. Mean length of hospitalization was 20 +/- 10 days (SD; median: 18 days). Closure of the colostomy using a small peristomal incision was performed in all surviving patients after a mean delay of 45 +/- 9 days (range 28-67 days). Mean length of hospitalization for colostomy closure was 7 +/- 3 days (range 3-18 days) without mortality. CONCLUSIONS: Applied systematically to all patients with diffuse peritonitis due to perforated diverticular disease, primary anastomosis was found to be as safe as the Hartmann procedure but appears to be superior in terms of total length of hospital stay, interval to stoma closure and rates of stoma closure. Primary anastomosis with diverting colostomy could constitute a valid alternative to the Hartmann procedure in selected patients with complicated diverticular disease, even in the presence of diffuse peritonitis.