Higher morbidity and mortality after combined total gastrectomy and pancreaticosplenectomy for gastric cancer.

Total gastrectomy with pancreaticosplenectomy for gastric cancer has been proposed for facilitating lymph node dissection or for resection of direct tumor invasion to the pancreas, especially for T4 lesions. Its effectiveness in improving patient survival is still controversial, and higher morbidity and mortality with this procedure have been reported in several series. Such risks to patient survival were not observed in the Japanese series. Based on a prospective gastric cancer database maintained from 1987 to 1999 in our institution, the morbidity and mortality were analyzed in our series of pancreaticosplenectomies. A total of 1,278 patients with gastric cancer received gastrectomy in our surgical unit. Of these, 127 patients underwent curative total gastrectomy with pancreaticosplenectomy in order to facilitate lymph node dissection or removal of direct tumor invasion. Operative time, postoperative hospital stay, postoperative complications, and surgical mortality were analyzed. Compared to another 201 total gastrectomies, longer mean operative time (7.91 +/- 2.16 hours vs. 6.67 +/- 2.01, p <0.001) and postoperative hospital stay (median, 24.5 days vs. 17, p <0.001) for combined organ resection (pancreaticosplenectomy) were shown in this series. The major complication rate, including intraabdominal abscess, anastomotic leak, postoperative bleeding, pancreatitis/fistula, chylous leak, and general complications causing unstable vital signs (26.8% vs. 11.9%, p = 0.001), but not the mortality rate (6.3% vs. 4.8%, p = 0.608), was also shown to be higher in pancreaticosplenectomy patients. The most frequent fatal complication was intraabdominal abscess. However, more than 50% of complications occurred in the first 40 pancreaticosplenectomies (1987-1991); after adequate accumulation of experience, the total complication rate (57.5% vs. 35.6%, p = 0.021), major complication rate (40% vs. 20.7%, p = 0.022), and mortality rate (17.5% vs. 1.1%, p = 0.001) improved significantly in the remaining 87 patients (1991-1999). We therefore conclude that total gastrectomy with pancreaticosplenectomy can be performed by experienced surgeons with acceptable risk of morbidity and mortality.