A series of 377 consecutive patients were operated upon with low anterior resection for rectal cancer in the nine Danish departments of surgical gastroenterology during 1992-1993. A retrospective analysis was carried out to calculate the frequency of anastomotic leakage and to evaluate factors of potential influence on the development of leakage according to the literature. Sixty-three patients (17%) developed leakage, which was followed by an increased mortality within the first three postoperative months. Only two variables significantly influenced the leakage rate: male gender was associated with a higher leakage rate (p = 0.02), whereas departments with a low number of rectal cancer surgeons had a low rate of anastomotic leakage (p = 0.02). In conclusion, the rather high frequency of anastomotic leakage calls for further clinical and pathogenetic research in this field. Until then, we recommend the routine use of a peroperative leakage test and selective use of prophylactic ostomy in cases of unsatisfactory anastomosis. Furthermore, it is recommended that low anterior resection for rectal cancer is limited to few surgeons in each department in order to ensure a uniform quality and hopefully also thereby reduce the rate of anastomotic leakage.