Risk Factors of Postoperative Pancreatic Fistula in Curative Gastric Cancer Surgery

ABSTRACT

PURPOSE: Postoperative pancreatic fistula is a dreadful complication after gastric cancer surgery. The purpose of this study is to evaluate the actual incidence and risk factors of postoperative pancreatic fistula after curative gastrectomy for gastric cancer. MATERIALS AND METHODS: A total of 900 patients who underwent gastrectomy for gastric cancer (laparoscopic gastrectomy, 594 patients; open gastrectomy 306 patients) were enrolled between January 2009 and December 2010. Clinical outcomes, including postoperative pancreatic fistula grade based on the International Study Group on Pancreatic Fistula, were investigated. RESULTS: Overall, the postoperative pancreatic fistula rate was 3.3% (30/900) (1.5% in laparoscopic gastrectomy versus 6.9% in open gastrectomy, P<0.001). Patients who underwent D2 lymphadenectomy, total gastrectomy, splenectomy or distal pancreatectomy showed higher postoperative pancreatic fistula rates (4.7%, 13.8%, 13.6%, or 57.1%, respectively, P<0.001). Patients with postoperative pancreatic fistula had higher morbidity (46.7% versus 13.1%, P<0.001), delayed gas out (4.9 days versus 3.8 days, P<0.001), belated diet start (5.8 days versus 3.5 days, P<0.001) and longer postoperative hospital stay (13.7 days versus 6.8 days, P<0.001). On the multivariate analysis, total gastrectomy (odds ratio 9.751, 95% confidence interval: 3.348 to 28.397, P<0.001), distal pancreatectomy (odds ratio 7.637, 95% confidence interval: 1.668 to 34.961, P=0.009) and open gastrectomy (odds ratio 2.934, 95% confidence interval: 1.100 to 7.826, P=0.032) were the independent risk factors of postoperative pancreatic fistula. CONCLUSIONS: Laparoscopic gastrectomy had an advantage over open gastrectomy in terms of the lower postoperative pancreatic fistula rate. Total gastrectomy and combined resection, such as distal pancreatectomy, should be performed carefully to minimize postoperative pancreatic fistula in gastric cancer surgery.