Despite improvement in surgical techniques and stapling devices during the last 10 years, colorectal anastomoses are still prone to leakage. The purpose of this study was to assess the performance and safety of stapled anastomoses in rectal surgery and to identify factors that influence the occurrence of anastomotic leaks. A review was undertaken of 1,014 patients who underwent stapled anastomoses to the rectum or anal canal for colorectal cancer or benign disease between 1989 and 1995 in a tertiary care institution. Indications for operations, comorbidities at admission, preoperative bowel preparation, stapler size, intraoperative events, associated surgical procedures, and clinical outcomes were tested for any association with anastomotic leak. A double stapled technique was used in 154 patients and a conventional single stapler technique was used in 860. Postoperative mortality was 1.6%, and the overall morbidity was 18.4%. Clinically apparent anastomotic leak developed in 29 patients (2.9%). Anastomotic dehiscence occurred in 22 of 284 patients (7.7%) after low stapling (within 7 cm from the anal verge) and in 7 of 730 patients (1%) after high stapling (p < 0.001). Diabetes mellitus, use of pelvic drainage, and duration of surgery were significantly related to the occurrence of anastomotic leak by the univariate analysis. Multivariate regression analysis identified an anastomotic distance from the anal verge within 7 cm as the only variable related to the occurrence of postoperative leak (p < 0.001). Low anastomoses were associated with a leak rate greater than with high colorectal anastomoses. We conclude that anastomoses to the rectum using the circular stapler can be done with low mortality and morbidity.