OBJECTIVE: The aim of this study was to assess if carbohydrate-deficient transferrin (CDT) and trypsin activity differentiate acute alcoholic pancreatitis from nonalcohol-related pancreatitis, and as a secondary goal to evaluate its use in comparison to healthy controls. METHODS: Serum levels of CDT and trypsin activity were measured in frozen sera from 70 nonconsecutive patients with acute pancreatitis and in 16 healthy controls. RESULTS: Causes of pancreatitis were gallstones in 51%, chronic alcoholism in 23%, and other or unknown causes in 26% of the patients. Serum CDT was significantly higher in alcoholic pancreatitis than in the nonalcoholic disease (p < 0.0001) with a median (interquartile range) of 30.8 U/L (23.6-41.7 U/L) in chronic alcoholism, 16.7 U/L (13.05-21.1 U/L) in gallstones, 17.5 U/L (15.9-21.6 U/L) in unknown cause, 19.3 U/L (15.1-27.7 U/L) in other etiologies, and 16.1 U/L (12.1-18.8 U/L) in controls. At a cutoff over 22.5 U/L, CDT showed a sensitivity of 87.5% and a specificity of 85.2%. Serum levels of trypsin activity were significantly higher (p = 0.0007) in alcoholic pancreatitis, median 165 U/L (76-405 U/L) than in nonalcoholic pancreatitis, median 73 U/L (46.5-100.5 U/L). At a cutoff value over 152 U/L, the sensitivity of trypsin activity was 60% with a specificity of 100%. In the multivariate analysis, patient's age (≤44 yr), CDT (>22.5 U/L), and trypsin activity (>152 U/L) enabled correct prediction of acute alcoholic pancreatitis in 98% of the cases. CONCLUSION: Serum CDT and trypsin activity are of clinical utility in differentiating alcoholic from nonalcoholic acute pancreatitis.