Objective: To answer the question whether a defunctioning stoma (DS) should be constructed routinely after total mesorectal excision or whether it could be used selectively to ensure patient safety.
Method: A PubMed search was performed. All randomized trials on the role of a DS were included. Also, observational articles published between January 1997 and August 2007 were reviewed. Sensitivity analysis of the mortality risk was performed.
Results: The clinical anastamotic leak (CAL) rate was 17% in 358 patients from four randomized trials and 9.6% in 4059 patients from 39 observational studies. The CAL rate increased significantly from 9.6% with DS to 24.4% without DS in four randomized trials, and from 7.9% with DS to 13.2% without DS in 17 observational studies. The re-operation rate as a result of anastomotic leakage was lower in patients with DS than in patients without DS in both study types. Leak-related mortality was not significantly different: 7.2% with vs 7.7% without DS in observational studies, and 0% with vs 4.6% without DS in randomized trials. Sensitivity analysis indicated that a selective DS strategy is acceptable if the CAL rate without DS is less than 16.6% with a CAL-related mortality of no more than 4.6%.
Conclusion: The results of this review support the routine construction of a protective stoma. However, selective use of a DS is justified from a patient safety point of view if the CAL-rate and its related mortality are limited. Each unit should audit its performance.