The value of serum Ca 19-9 dosage for pancreatic carcinoma diagnosis has been studied in heterogeneous series. The effect of the complications of chronic pancreatitis and pancreatic carcinoma on serum Ca 19-9 value has not been assessed precisely. The aims of this study were to assess: a) the value of Ca 19-9 to differentiate benign from malignant pancreatic disease; b) the influence of complications (particularly, cholestasis).The studied population included 179 patients: 126 with chronic pancreatitis (25 females, 101 males, 45 with cholestasis) and 53 with pancreatic carcinoma (27 females, 26 males, 37 with cholestasis).At 37 UI/mL threshold, the specificity and sensitivity of Ca 19-9 were 53 and 95%, respectively. Cholestasis was associated with a significant increase of Ca 19-9 in patients with chronic pancreatitis but not in those with pancreatic carcinoma. At 300 UI/mL threshold, the specificity and sensitivity of Ca 19-9 were 95 and 81% in patients without cholestasis and 87 and 81% in those with cholestasis, respectively. Diabetes mellitus was associated with a significant increase of Ca 19-9 only in patients with chronic pancreatitis without cholestasis. Pancreatic calcifications, pseudocysts, cirrhosis, pleural effusion or ascites were not associated with significant variation of Ca 19-9.In patients with pancreatic disease, 300 UI/mL threshold is the most accurate to differentiate benign from malignant disease, whatever the presence of cholestasis.