Pancreaticoduodenectomy for cancer of the head of the pancreas

Many modifications of the original technique have been proposed to try to improve the results of the pancreaticoduodenectomy described by Whipple. To evaluate these modifications, we have reviewed randomized controlled trials, meta-analyses, and well-conducted retrospective series. Neither total pancreatectomy nor extended lymph node dissections have shown an improval of survival in retrospective studies. Preoperative evidence of mesenteric or portal vein involvement does not contraindicate pancreatic resection and survival rates are similar to those of patients with no venous involvement. Prospective trials and one meta-analysis have shown neither advantage nor disadvantage of pylorus-preserving pancreaticoduodenectomy or of pancreatico-gastric anastomosis. Three trials and one meta-analysis of pancreatico-gastric anastomosis have failed to demonstrate a decrease in the risk of pancreatic fistula. Two trials suggest that the risk of fistula formation is decreased by implantation of the pancreatic remnant into die jejunum or by trans-jejunal stenting of the pancreaticojejunal anastomosis with external drainage; but these findings are not supported by a third trial. The results of the antisecretory use of somatostatin are contradictory. Leak and fistula formation were decreased when the criteria for leakage was based on laboratory findings; but in 4 out of 5 trials, somatostatin did not decrease the incidence of clinical fistula. The use of fibrin glue to occlude the pancreatic duct or seal the cut surface of the pancreas did not decrease the rate of intra-abdominal complications. In conclusion, the pancreaticoduodenal resection described by Whipple may still be considered the gold standard for resection of pancreatic cancer. The technical experience of surgeons and their institutional support staff resulted in lower perioperative morbidity and mortality and in higher survival rates.