The aim of this retrospective study was to analyze the risk factors for pancreatic anastomotic leakage after pancreatoduodenectomy (PD) and to determine whether duct-to-mucosa pancreaticojejunostomy is superior to the total external tube drainage technique. Between 1990 and 1999, 161 patients underwent PD with end-to-side pancreaticojejunostomy at our institution. Fourteen preoperative and ten intraoperative risk factors for pancreaticojejunal anastomotic leakage were analyzed. Pancreaticojejunal anastomotic leakage was identified in 11% (17/161) of the patients. No preoperative parameters were found to have a significant association with the risk of pancreatic leakage. Three intraoperative parameters were identified as significant by means of univariate analysis: anastomotic technique, pancreatic duct size and texture of the remnant pancreas. A duct-to-mucosa pancreaticojejunostomy with total external tube drainage (3% vs. 15%, p = 0.018). A pancreas without duct dilatation of soft pancreas was more likely to develop pancreatic leakage than one with duct dilatation or atrophy. A multivariate analysis revealed that only anastomotic technique turned out to be an independent risk factor (Odds ratio: 4.15, CI: 1.1-27.4). Sub-analysis of patients with soft pancreas and non-dilated pancreatic duct further supported the finding that the duct-to-mucosa pancreaticojejunostomy technique is safer for patients at high risk. Results indicate that the status of the remnant pancreas and the pancreaticojejunostomy technique are the substantial risk factors for pancreatic leakage after pancreatoduodenecomy. Duct-to-mucosa pancreaticojejunostomy might well be the procedure of choice.