Analytical performance of two enzymatic methods for glycated albumin

Background: The measurement of glycemic control among patients with diabetes mellitus is important for predict-ing the risk of diabetic complications. Glycated hemoglobin (HbA1c) measurements have been used for long-term glycemic control in clinical practice. However, glycated albumin (GA) or glycated serum protein (GSP) is a more reliable indicator of glycemic control in the short term (2 - 4 weeks) and an alternative marker of HbA1c in clini-cal situations with changing red blood cell (RBC) lifespan. Here, we evaluated an analytical performance of the two enzymatic assays commercially available, Lucica GA-L and Autolab GA, for the determination of GA (%). Methods: For each assay, the imprecision was evaluated based on CLSI EP05-A2. In total, serum samples of 283 subjects were simultaneously tested using the two enzymatic assays for method comparison according to CLSI EP09-A3. Some subjects collected the laboratory data for HbA1c. Results: The GA (%) value of the Lucica GA-L assay showed highly reproducible results with within-run, be-tween-run, and total coefficient of variations (CVs) below 2.4%. The Autolab GA assay also showed reliable re-sults with within-run, between-run, and total CVs below 3.9%. The Lucica GA-L assay showed a very high corre-lation with the Autolab GA assay (r = 0.9993). However, at the median decision point (MDP, 14.3%), the estimat-ed bias of the Autolab GA assay was 4.5%, exceeding the allowable bias (2.9%) accounting for the biological vari-ation. For the correlation analysis between HbA1c and GA (%), the two assays demonstrated the same pattern, with no statistical differences between the two independent correlation coefficients. Conclusions: Both GA assays evaluated in this study showed good precision and excellent correlation, but the comparability at MDP did not meet the acceptance criteria.