The most feared complication of gastrointestinal tract operations is anastomotic leakage, not only because of the presumed individual surgeon's culpability but also because of the assumption that this event is often fatal. We have experienced 32 cases of anastomotic leakage after elective gastric resection during 8 years. The purpose of this study was to evaluate the result of their treatment.

We evaluated the records of 1335 patients who had undergone elective gastric resection for an adenocarcinoma of stomach from January 1995 to October 2003 and conducted a retrospective, multivariate analysis.

Of the 1335 patients, 32 (2.4%) sustained an anastomotic leakage. Anastomotic leakages usually developed on mean postoperative day 9.1±3.2 (range:1~18 days). Overall, 31.3% (10/32) of patients who sustained an anastomotic leakage died. The anastomotic leakages were identifed by radiological study or by operative finding at the site of the duodenal stump (20 patients), the esophagojejunostomy (7), the gastroduodenostomy (4), and the gastrojejunostomy (1). Fourteen patients (43.8%) underwent a relaparotomy, a drainage procedure in the main, and 18 patients (56.3%) were treated conservatively. The mortality rates were 42.9% (6/14) and 22.2% (4/18), respectively, but this difference was not statistically significant. A cox's proportional hazard analysis showed that a body-mass Index < 24 kg/m² (odds ratio 5.55, 95% CI: 0.69~44.82) and non-enteral feeding (odds ratio 18.27, 95% CI 2.22~150.69) were independent factors of mortality due to anastomotic leakage.

Our observations show that anastomotic leakage after an elective gastric resection has a high risk of being fatal. Moreover, for a patient with a body-mass index lower than 24 kg/m± and/or non-enteral feeding, an anastomotic leakage after an elective gastric resection has a higher risk of being fatal.