The consistency of aminotransferase analysis in China: a comparison of six mainstream aminotransferase routine methods and recalibration using human pooled serum preparations supplemented with human recombinant aminotransferases

Background: To evaluate the consistency of six mainstream homogeneous systems for aminotransferase measurements and improve the consistency of measurements by applying uniform calibrators.Methods: 200 individual samples were grouped into four sets for assays with and without pyridoxal-5-phosphate (P-5'-P) for Alanine aminotransferase (ALT) and aspartate aminotransferase (AST). Routine assays for the P-5'-P group were compared with a reference measurement procedure (RMP). In the non-P-5'-P group, four routine assays were analyzed in a pairwise method across six method pairs. Recalibration was performed using human serum pools (HSPs) supplemented with human original recombinant aminotransferases (HOR). Data were analyzed by Passing-Bablok regression and Bland-Altman plots.Results: In the P-5'-P group, the mean biases for Ortho and Dimension assays for ALT were 17.0% and -25.4%, respectively; for AST, the mean biases were -9.5% and -9.6%, respectively. In the non-P-5'-P group, the mean deviations ranged from -5.9% to 5.9% for ALT. For AST, the relative deviations ranged from -19.1% to 6.5%. After recalibration, in the P-5'-P group, the relative biases were reduced to -12.2% to 7.7% for ALT and -6.9% to 0.8% for AST. The mean deviations for the non-P-5'-P AST group were reduced remarkably (-3.0% to 3.3%).Conclusion: Assays supplemented with P-5'-P exhibited poor performance against RMP for both ALT and AST. For assays without P-5'-P, AST results showed non-satisfactory comparability for almost all method pairs. Uniform calibrators such as HSPs supplemented with HOR could improve consistency among the mainstream homogeneous systems for the measurement of aminotransferase activity, particularly for AST measurement.