The aim of this study was to determine whether leakage from a colorectal anastomosis following potentially curative anterior resection for rectal cancer is an independent risk factor for local recurrence. The study included all patients who had a potentially curative anterior resection with anastomosis for adenocarcinoma of the rectum between 1971 and 1991 at Concord Hospital. The data were collected prospectively, with complete follow-up for at least 5 years. The Kaplan-Meier method was used to compare time to recurrence between strata of categorical variables. Proportional hazards regression was used in multivariate modelling. There were 403 patients in the study. After adjustment for lymph node metastases, the distal resection margin of resection, non-total anatomical dissection of the rectum and the level of anastomosis, multivariate analysis identified a significant association between anastomotic leakage and local recurrence (hazard ratio 3.8, 95 per cent confidence interval 1.8 to 7.9). Leakage following a colorectal anastomosis after potentially curative resection for adenocarcinoma of the rectum is an independent predictor of local recurrence. Copyright 2003 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.