A systematic search was performed of the published literature. Data on the definition and incidence rate of AL, postoperative mortality caused by AL, and overall postoperative mortality were extracted. Data were pooled and a meta-analysis was performed.
Twenty-two studies with 10,343 patients in total were analyzed. Meta-analysis of the data showed an average AL rate of 9 % , postoperative mortality caused by leakage of 0.7 % and overall postoperative mortality of 2 % . The studies showed variation in incidence, definition and measurement of all outcomes.
We found a considerable overall AL rate and a large contribution of AL to the overall postoperative mortality. The variability of definitions and measurement of AL, postoperative mortality caused by leakage and overall postoperative mortality may hinder providing reliable risk information. Large-scale audit programs may provide accurate and valid risk information which can be used for preoperative decision making.