RESUMO
BACKGROUND: Recently, a new EIA method for pancreatic amylase was introduced that was assayed by inhibition of the salivary amylase using the synergistic action of two monoclonal antibodies. We evaluated the clinical usefulness of the pancreatic amylase by using the sensitivity, the specificity and diagnostic accuracy of the receiver-operator characteristics (ROC) curve. METHODS: We divided into 3 groups: acute pancreatitis (n=26) diagnosed by ultrasonography and computed tomography, control patients (n=105), and healthy controls (n=95). Serum total amylase, pancreatic amylase, and lipase were assayed by the Hitachi 7170. The upper limit of the reference range of the total amylase, pancreatic amylase, and lipase was respectively 216 U/L, 115 U/L and 200 U/L in this hospital. RESULTS: The sensitivity of total amylase, pancreatic amylase, and lipase for the diagnosis of acute pancreatitis was 73.1%, 88.5%, and 92.3%, respectively. The specificity of total amylase, pancreatic amylase, and lipase was 70.5%, 81.9%, and 82.9%, respectively. The diagnostic accuracy, determined as the area under the curve, was 0.795 for total amylase, 0.868 for pancreatic amylase, and 0.886 for lipase. There was a significant difference between the total amylase and pancreatic amylase (P=0.045), but not a significant difference between the pancreatic amylase and lipase (P=0.613) by ROC curve. CONCLUSIONS: Pancreatic amylase had a higher sensitivity, specificity, and diagnostic accuracy than the total amylase, and showed a similar diagnostic performance as lipase. Therefore, we concluded that the pancreatic amylase was a better diagnostic tool than the total amylase in the diagnosis of acute pancreatitis.