Operative mortality rates after pancreaticoduodenectomy (PD) have decreased over the past 3 decades and recent series have reported very low mortality. Nevertheless pancreatic fistula remains the major cause of morbidity. A significant fraction of patients undergoing PD develop a postoperative pancreaticocutaneous leak. To analyze the clinical application of risk factors for pancreatic leakage after PD, we conducted a retrospective review of patients undergoing PD. METHODS: Thirty-seven patients who underwent PD at our hospital between January 2004 and July 2005 were reviewed retrospectively. The primary diseases of the patients included pancreas cancer, ampullary cancer, bile duct cancer, islet cell cancer, duodenal cancer, chronic pancreatitis, pancreatic cystadenoma, gastric cancer, and pancreatic metastases. Standard PD was performed for 3 cases and pylorus-preserving PD for 34 cases. A duct-to-mucosa pancreaticojejunostomy was performed in all patients and a transanastomotic stent was placed in patients with soft pancreas and non-dilated duct. Octreotide was not used after surgery. Four preoperative, two intraoperative, and five postoperative risk factors with potential to affect the incidence of pancreatic fistula were analyzed. The recent established definition of pancreatic fistula was used to make the fistula diagnostic - a drain output of any measurable volume of fluid on or after postoperative day 3 with an amylase content greater than 3 times the serum amylase. RESULTS: Of the 37 patients, 11 (29.7%) were identified as having pancreatic leakage after operation. The hospital mortality in this series was 5.4% (2/37), and the mortality associated with pancreatic fistula was 0% (0/37). General risk factors including patient age, gender, history of jaundice, preoperative nutrition, and pathological diagnosis didn’t have any relation with pancreatic fistula occurrence. Intraoperative risk factor, texture of the remnant pancreas, was not found to be significantly associated with pancreatic leakage. Blood loss, type of resection, and serum amylase level on postoperative day 1 did not have significantly relation with pancreatic fistula. The incidence of pancreatic leakage was 81.9% (9/11) in those patients with abdominal drain amylase level > 1000UI/dL on postoperative day 1, it was the only risk factor with significantly relation with pancreatic fistula (p<0.05). CONCLUSION: Amylase level > 1000UI/dL in the abdominal drain on postoperative day 1 shows a significant relation with the occurrence of pancreatic fistula and seems to be an important factor for early identification of development of pancreatic fistula after PD.