Lin Xi-Zhang M.D.; Wang, Sun-Sang M.D.; Tsai, Yang-Te M.D.; Lee, Shou-Dong M.D.; Shiesh, Shu-Chu M.S.; Pan, Huay-Ben M.D.; Su, Cheng-Hsi M.D.; Lin, Ching-Yih M.D.Journal of Clinical Gastroenterology: February 1989 Clinical Studies: PDF Only Abstract We evaluated the diagnostic value of serum amylase, isoamylase, and lipase for the diagnosis of acute pancreatitis from sera of patients with acute abdominal pain. Comparison was first made in condition A between 32 patients with image-proven pancreatitis and 414 patients with nonpancreatic acute abdomen (the control group), then in condition B, between 62 pancreatitis patients with or without image proof and the control group. We found (a) that patients with image-proven pancreatitis suffer a more severe clinical course than those without; (b) that the sensitivity, positive predictive value, and accuracy in condition B are higher than in condition A at any cutoff level; (c) that none of the enzyme assays is specific at the upper reference limit, but their diagnostic yields are much improved by raising cutoff levels to about three or four times the upper limit; and (d) that at these selected cutoff levels, amylase had a diagnostic value similar to p-isoamylase or lipase in both conditions (sensitivity 84% and 92% for amylase in conditions A and B, respectively; specificity 98% and 98%; positive predictive value 75% and 90%; negative predictive value 99% and 99%; accuracy 91% and 97%). In conclusion, at an appropriately selected cutoff level, amylase can be effectively used as the first-line test and isoamylase or lipase as adjunct tests for acute abdominal conditions. © Lippincott-Raven Publishers.