Résumé : The sensitivity and specificity of amylasemia, the ratios of amylase/creatinine clearance and amylasuria/creatininuria were determined in four groups of patients: a control group (n = 43), patients with acute pancreatitis detected on computed tomography (n = 30, 25 cases of alcoholic pancreatitis), patients with an acute surgical abdomen without pancreatitis (n = 25), and patients with renal failure (n = 20). Sensitivity was defined for the acute pancreatitis group and specificity for the other groups. When amylasemia was > 20 UI/dl and the amylasuria/creatininuria ratio > 100, sensitivity was 98 per cent. The specificity of these two results in patients with an acute surgical abdomen was 98 pe cent. When the ratio amylase/creatinine clearance ratio was > 4 sensitivity was 73 per cent and specificity in patients with acute surgical abdomen was 75 per cent. These two values were lower than those of the two preceeding tests (p < 0.01). Sensitivity of the association of an amylasemia greater than 13 UI/dl (m + 25D) with a clearance ratio > 4 was 73 per cent. The amylase/creatinine clearance ratio did not seem to be reliable since its change was delayed with respect to the increase of amylasemia and amylasuria. This ratio has a poor specificity as it increased when the clearance of creatinine decreased in the group with an acute surgical abdomen associated with functional or organic renal failure. In these two groups, the correlation between the amylase/creatinine clearance ratio and creatininemia was significant. This suggested that the clearance of creatinine fell more rapidly than the clearance of amylase as renal failure increased.