Impact of anastomotic leakage on long-term survival after total gastrectomy for carcinoma of the stomach.

BACKGROUND: Recent studies suggest that anastomotic leak may adversely affect long-term survival in patients undergoing surgery for gastrointestinal malignancies. Data relating to total gastrectomy for gastric cancer are scarce.

METHODS: An electronic database of all patients with resectable gastric cancer treated between January 1999 and December 2004 at seven university surgical centres cooperating in the Polish Gastric Cancer Study Group was reviewed.

RESULTS: Anastomotic leakage was diagnosed in 41 (5.9 per cent) of 690 patients who underwent total gastrectomy. The prevalence of surgical and general complications, and mortality rates were significantly higher in patients diagnosed with anastomotic leakage. The only two independent risk factors for leakage were Eastern Cooperative Oncology Group performance status of 2 or 3 (odds ratio 5.09, 95 per cent confidence interval (c.i.) 2.29 to 11.32) and splenectomy (odds ratio 2.58, 95 per cent c.i. 1.08 to 6.13). Two Cox proportional hazards models including all the patients and excluding in-hospital deaths identified anastomotic leakage as an independent predictor of survival with hazard ratios of 3.47 (95 per cent c.i. 1.82 to 6.64) and 3.14 (1.51-6.53) respectively.

CONCLUSION: The occurrence of anastomotic leakage was a major independent prognostic factor for long-term survival.