OBJECTIVE:

To determine which biochemical test is best to distinguish acute pancreatitis from other pancreatic and nonpancreatic diseases associated with hyperamylasemia.

DESIGN:

We conducted a prospective clinical study of 836 consecutive patients who had a total serum amylase test requested by a physician during a 7-month period.

MATERIAL AND METHODS:

Radioimmunoassay and enzymatic activity methods were used to measure pancreas-specific proteins of varied size, charge, and stability. In addition, scoring systems were used for the diagnosis of pancreatitis, and statistical analyses were done to determine sensitivity and specificity.

RESULTS:

We found minor differences in sensitivity and specificity for diagnosis of acute pancreatitis among pancreatic isoamylase, phospholipase A2, colipase, lipase, and carboxylester lipase. Of these tests, the combination of isoamylase and phospholipase A2 had a small but statistically significant increased sensitivity (90%; 95% confidence interval [CI] = 74 to 98%) and specificity (93%; 95% CI = 91 to 95%) over isoamylase (90% and 92%, respectively; 95% CI = 90 to 94%) and phospholipase A2 (90% and 75%, respectively; 95% CI = 72 to 78%) alone for the diagnosis of acute pancreatitis.

CONCLUSION:

Pancreas-specific proteins are satisfactory for diagnosing acute pancreatitis if the test is validated by the laboratory. Clinically, the slight advantage of using both isoamylase and phospholipase A2 does not outweigh the expense of performing two assays; we recommend using isoamylase to diagnose acute pancreatitis.