Characterization and mathematical correction of hemolysis interference in selected Hitachi 717 assays.

The effect of hemolysis on several assays performed with the Hitachi 717 was quantified by relating the amount of error to the concentration of hemoglobin. Hemolysis interference was judged clinically significant when analyte concentration varied by > 10% from the initial value. Hemolysis interference was significant for alkaline phosphatase, aspartate aminotransferase, alpha-amylase, bilirubin, creatine kinase, gamma-glutamyltransferase, lactate dehydrogenase, lactate dehydrogenase-1, potassium, and theophylline assays. Error (expressed in absolute terms) was linearly dependent on hemoglobin concentration and independent of the initial analyte concentration in each case, except for bilirubin and theophylline, where multiple regression analysis was required to quantify the effect. Relative error was dependent on the initial analyte concentration in all cases. Correction formulas were calculated from linear regression of absolute error vs hemoglobin concentration. Clinical application of correction formulas and mechanisms of hemolysis interference for each assay are discussed.