Pancreaticoduodenectomy with or without distal gastrectomy and extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma--part 3: update on 5-year survival.

The study objective was to update the survival analysis at the 5-year mark of patients undergoing standard versus radical (extended) pancreaticoduodenectomy (PD) for periampullary adenocarcinoma (cancers of the pancreas, ampulla, common bile duct, and duodenum). A prospective randomized trial was performed (April 1996 through June 2001) comparing survival after pylorus-preserving PD resection (standard) to survival after PD with distal gastrectomy and retroperitoneal lymphadenectomy (radical). An interim report (Ann Surg 1999;229:613) and report after closing the trial (Ann Surg 2002;236:355) showed no differences in survival between the standard and radical groups. Two hundred ninety-nine patients were randomized to either the standard or radical group. Five patients were excluded from final analysis because final pathology failed to reveal adenocarcinoma. The 5-year survival of the two groups was evaluated. The median live patient follow-up is now 64 months (5.33 years). For all periampullary cancer patients, those undergoing standard resection had 1- and 5-year survival rates of 78% and 25%, respectively, compared with 76% and 31% (P = 0.57) for those patients in the radical group. For pancreatic adenocarcinoma patients, the 1- and 5-year survival rates in the standard group were 75% and 13%, respectively, compared with 73% and 29% in the radical group (P = 0.13). The increased morbidity rate, longer operative time, and similar survival for radical PD led us to conclude that pylorus-preserving PD without retroperitoneal lymphadenectomy should be the procedure of choice for most patients with resectable periampullary adenocarcinoma. While there is an intriguing trend toward improved survival in patients with pancreatic adenocarcinoma in the radical group, this trend may be largely accounted for by the higher incidence of microscopically margin positive resections in the standard resection group (21%) compared with a 5% incidence in the radical group (P = 0.002).