Diagnostic Assays in Acute Pancreatitis

A Study of Sensitivity and Specificity

  1. WILLIAM M. STEINBERG, M.D.;
  2. STAFFORD S. GOLDSTEIN, M.D.;
  3. NANCY D. DAVIS, P.A.;
  4. JOHN SHAMMA'A, M.D.; and
  5. KIMBERLY ANDERSON, B.S.
  1. Washington, D.C.

    Abstract

    The sensitivity and specificity of five assays used to diagnose acute pancreatitis were studied: two amylase assays; one lipase; one trypsinogen; and one pancreatic isoamylase. Thirty-nine patients with acute pancreatitis were compared to 127 controls with abdominal pain. Using the upper limit of normal both amylase assays appeared sensitive but somewhat nonspecific (specificities of 88.9% and 86%, respectively). The trypsinogen and pancreatic isoamylase assays were also relatively nonspecific (specificity of 82.8% and 85.1%). Most nonspecific elevations occurred between a one- and twofold elevation of each assay. Lipase, however, maintained excellent specificity (99%) at its upper limit of normal. If the level of best cutoff is used instead (the level that best enhances sensitivity and specificity), the specificities of both amylase assays, as well as the trypsinogen and pancreatic isoamylase assays, exceed 95%. At the best cutoff level, trypsinogen maintains a qualitative advantage in sensitivity over lipase or pancreatic isoamylase (97.4% as compared to 86.5% and 84.6%).

    Article and Author Information

    • ▸From the Department of Medicine, Division of Gastroenterology, The George Washington University Medical Center; Washington, D.C.

    • ▸Requests for reprints should be addressed to William M. Steinberg, M.D.; Division of Gastroenterology, 2150 Pennsylvania Avenue, N.W.; Washington, DC 20037.

    « Previous | Next Article »Table of Contents