OBJECTIVE: The objective of this study was to elucidate the significance and limitations of serum amylase levels in the diagnosis of blunt injury to the pancreas. SUMMARY BACKGROUND DATA: Several recently published reports of analyses of patients with blunt abdominal trauma have indicated that determination of the serum amylase level on admission seemed to be of little value in the diagnosis of acute injury to the pancreas. Few previous reports have described clearly the significance and the limitations of the serum amylase level in diagnosing injury to the pancreas. METHODS: Retrospective analysis of 73 patients with blunt injury to the pancreas during 16-year period from February 1980 to January 1996 was performed. The factors analyzed in the current study included age, gender, time elapsed from injury to admission, hypotension on admission, type of injury to the pancreas, intra-abdominal- and intracranial-associated injuries, and death. RESULTS: The serum amylase level was found to be abnormal in all patients admitted more than 3 hours after trauma. Various comparisons between patients with elevated (n = 61, 83.6%) and nonelevated (n = 12, 16.4%) serum amylase levels showed the statistical significance solely of the time elapsed from injury to admission (7 +/- 1.5 hours vs. 1.3 +/- 0.2 hour, p < 0.001). The major factor that influences the serum amylase level on admission appeared to be the time elapsed from injury to admission. Determination of the serum amylase level is not diagnostic within 3 hours or fewer after trauma, irrespective of the type of injury. CONCLUSIONS: To avoid failure in the detection of pancreatic injury, the authors advocate determination of serum amylase levels more than 3 hours after trauma.