We report the results of the analytical and clinical evaluation of a specific enzyme immunoassay for determination of human pancreatic lipase, in comparison with a turbidimetric method, in pancreatic pathology. Under standardized conditions of incubation time and temperature we found intraassay coefficient of variation (CV) of 1.3, 3.2, 2.1% at means = 18.7, 43.7, 224 micrograms/L and interassay CV of 2.8, 4.6, 3.0% at means = 19, 42.6, 230 micrograms/L, respectively. In general, a good correlation (r = 0.97) was found between lipase determined as a protein or through its catalytic activity. No significant correlation (r = 0.38) was observed with samples containing low concentration of lipase (up to 18 micrograms/L). We conclude that the turbidimetric method is reliable for routine determinations in the diagnosis of acute pancreatic pathology. However, the better sensitivity of the immunochemical assay should provide additional information for monitoring pancreatic insufficiency.