A randomised study of colostomies in low colorectal anastomoses.

Objective: To assess the value of covering colostomy for patients undergoing low anterior resection for rectal neoplasms.

Design: Prospective randomised study.

Setting: Two university hospitals, Finland.

Subjects: 38 patients with air-tight stapled end-to-end anastomoses and complete anastomotic tissue rings were randomly allocated to have a covering colostomy (n = 19) or not.

Main outcome measures: Postoperative mortality, anastomotic leaks, reoperations for leaks.

Results: The clinical leak rate was 24% (9/38) and six patients (16%) had radiological leaks. The total number of leaks (clinical and radiological together) was similar in the two groups, 7/19 compared with 8/19, respectively. There were fewer clinical leaks in the colostomy group (3/19; 16% compared with 6/19; 32%), but the difference was not significant. Reoperations for leaks were necessary more often in patients who did not have a covering colostomy (6/19; 32% compared with 1/19; 5%, p = 0.09). Two patients who did not have a stoma died from the infective complications of their leaks and one died of heart failure in the colostomy group. One patient who had not been given a stoma initially was left with a permanent colostomy after a leak.

Conclusions: Our results suggest that a covering colostomy does not reduce the leak rate after low anterior resection, but prevents most of the severe infective consequences of the leaks.