Ten clinical centres in Italy participated in an evaluation of a pancreas-specific
alpha-amylase assay using two
monoclonal antibodies. Comparisons with electrophoretic methods showed good agreement in the reference range, but systematic deviations above it. The diagnostic information of the two methods appeared substantially different if the percentage values from electrophoresis were compared with activity units from the immunoinhibition method, but became similar if the two methods were compared on the basis of activity units. Reference intervals determined for serum/plasma corresponded to those previously published, but those determined for urine differed slightly from the published values. Clinical sensitivities for the assessment of
acute pancreatitis (n = 134) were found as follows. Related to the upper limit of the control group (n = 141), the
pancreatic alpha-amylase, total
alpha-amylase and
lipase showed sensitivities of 0.94, 0.87 and 0.93, respectively. When the cut-off point was set at the three-fold upper limit of the control group, sensitivities of 0.73, 0.53 and 0.70 were found, and the specificity was 1.00 for all three methods. Based on this commonly used higher cut-off point, the determination of
lipase in addition to
pancreatic alpha-amylase enhanced the sensitivity in the recognition of
acute pancreatitis by 8%; conversely, the determination of
pancreatic alpha-amylase in addition to
lipase increased the number of true positive results by 13%. The high practicability and interlaboratory transferability documented in the results of the collaborative study show that the
pancreatic alpha-amylase assay is very useful for recognizing pancreatic
inflammations, especially in combination with
lipase.