This study tried to determine if drainage fluid amylase reflects pancreatic leakage after pancreaticoduodenectomy and to determine the factors affecting the drainage amylase level. Patients undergoing pancreaticoduodenectomy were recruited. The drainage amylase was measured from postoperative day (POD) 1 to POD 7. Direct evidence of pancreatic leakage was provided by upper gastrointestinal studies using a water-soluble contrast medium and methylene blue dye in the pancreaticogastrostomy group or by pancreaticography with injected contrast medium via an exteriorized pancreatic stent in the pancreaticojejunostomy group on POD 7. A total of 37 patients were recruited. The drainage amylase level was higher than the normal serum amylase (>or= 190 U/L) in more than half of the cases on the initial POD 2 specimen, with a median of 745 U/L on POD 1 and 663 U/L on POD 2. The drainage amylase level was more than three times the normal serum amylase level (>or= 190 x 3 U/L) in 56.8% on POD 1, in 51.4% on POD 2, and in nearly one-third on POD 7 (29.7%). However, no pancreatic leakage occurred in any of the patients with a drainage amylase of >or= 190 U/L. Only one case of pancreatic leakage with a small amount of drainage fluid (10 ml) and low amylase level (74 U/L), was noted. Soft pancreatic parenchyma and a nondilated pancreatic duct were significantly associated with higher drainage amylase levels. In conclusion, biochemical leakage defined by amylase-rich drainage fluid might have no clinical significance and was not necessarily clinical pancreatic leakage following pancreaticoduodenectomy.