Development of Stabilizing Hemoglobin A2 Reference Materials and Assessment of Commutability using Multiple Statistical Models

Abstract

One of the challenges in developing hemoglobin A2 (HbA2) reference materials is achieving reliable stability, homogeneity and value transfer capabilities. Another challenge lies in evaluating commutability and deciding which statistical methods to use for the data while ensuring statistical power and avoiding the use of unnecessary statistics. In this study, we developed two levels of HbA2 reference materials and assessed commutability using multiple statistical models. We developed a new combined formulation buffer to maintain homogeneity and stability. Two medical concentrations of HbA2 reference materials were prepared. According to the guidelines of the Clinical Laboratory Standards Institute (CLSI), we determined the reference values of the HbA2 reference materials and evaluated their homogeneity, stability and value transfer. Furthermore, the commutability of the HbA2 reference materials was evaluated using five routine analyzers. Four statistical methods were applied to assess commutability, and the strengths and weaknesses of the different statistical models were described. The HbA2 reference materials were 2.13% ± 0.21%, k = 2 and 6.23% ± 0.48%, k = 2, respectively. Furthermore, the reference materials showed good homogeneity, long-term and short-term stability, reliable value transfer and compatibility with all the routine analyzers. Passing–Bablok regression analysis using the four statistical methods was found to be the better method for evaluating the commutability of the reference materials when uncertainty was disregarded or not considered. The HbA2 reference materials demonstrated good homogeneity, stability, and reliability in terms of value transfer and commutability. Furthermore, Passing–Bablok regression was a more suitable method for evaluating commutability.