Diabetes and risk of anastomotic leakage after gastrointestinal surgery.

ABSTRACT

BACKGROUND: Anastomotic leakage (AL) is one of the most common and lethal complications in gastrointestinal surgery. However, the relationship between AL risk and diabetes mellitus (DM) remains ambiguous. This meta-analysis was to evaluate the association between DM and AL risk in patients after gastrointestinal resection. METHODS: Odds ratios (OR) estimate with their corresponding 95% confidence intervals (CIs) were combined and weighted to produce pooled OR using the fixed-effects model. Relative risks were calculated in subgroup analysis of prospective studies. We calculated publication bias by Begg rank correlation test and Egger linear regression test. RESULTS: DM was significantly and independently associated with an increased risk of AL morbidity in colorectal patients, 1.661 times in total patients (95% CIs = 1.266-2.178), 1.995 times in a subgroup of case-control studies, 1.581 times in cohort investigations, 1.688 times in retrospective trials, and 1.562 times in prospective designs. After adjusting for the factor of obesity and/or body mass index in the subgroup analyses of colorectal surgery, DM patients without obesity experienced a significantly increased risk of AL (OR = 1.572, 95% CIs = 1.112-2.222). Furthermore, when obesity had not been adjusted, DM patients endured a dramatical increase of AL incidence (OR = 1.812, 95% CIs = 1.171-2.804). Perforation incidence after gastric resection showed borderline association with DM (OR = 2.170, 95% CIs = 0.956-4.926). CONCLUSIONS: The present meta-analysis provides strong evidence for the first time that DM is significantly and independently associated with an increased risk of AL mortality in colorectal surgery.