Introduction: Hemoglobin A1c (HbA1c) measurement devices are widely used to evaluate glycemic control in diabetic patients. The aim of this study was to investigate the comparability of various HbA1c instruments used in Iran.
Methods: In the present study, 154 fresh whole blood samples from diabetic patients, with different HbA1c levels (4.0%-10%) and no types of hemoglobinopathy were analyzed by six HbA1c assays including one high performance liquid chromatography (HPLC) method (D10 HbA1c), two immunoassay methods (COBAS INTEGRA 400 and Pars Azmoon kit), one Boronate affinity method (Nycocard Reader II), and two ion exchange methods (Biosystems and DS5). The two National Glycohemoglobin Standardization Programs (NGSP) certified system, D10 and COBAS INTEGRA 400 which are certified as secondary reference measurement procedures, were considered as reference methods. The CLSI document (EP9-A2) - Method comparison and Bias estimation using patient samples, approved guideline - was used to compare the performance of different HbA1c instruments.
Results: The mean of HbA1c in all four types of assays was less than the reference methods (P-value < 0.01).The mean of absolute difference between the reference methods was the least (0.11%). Among the other four tests, Biosystems had the smallest mean of difference (-0.21%), while Pars Azmoon had the highest (-1.18%). Pars Azmoon showed the greatest difference (95% confidence interval) when compared to D10 [-15.5%(-5.7%to -25.3%)] and COBAS INTEGRA [-17% (-9.16% to -24.84%)]. The highest regression slope (B) was found in DS5 method (0.96) in regression model with both reference methods.
Conclusion: It can be concluded that although HbA1c standardization programs have resulted in great improvements in the comparability of HbA1c assays, unacceptable errors still exist and further national and international projects are required for standardization of HbA1c measurement. In this situation, it is recommended to use the same laboratory for HbA1c measurement to monitor diabetic patients.