RESUMO
BACKGROUND AND OBJECTIVES: Hemoglobin A1c (HbA1c) is widely used for the monitoring and management of diabetes mellitus. The aim of this study is to investigate the influence of hemoglobin (Hb) variants on the measurement of HbA1c. MATERIALS AND METHODS: HbA1c levels of 845 blood samples obtained from diabetic patients with various hemoglobin types were measured using a turbidimetric inhibition immunoassay and capillary electrophoresis. RESULTS: Of 845 patients with diabetes, 65.7% (555/845) have the normal hemoglobin type (A2A) and 34.3% (290/845) have various abnormal hemoglobin types, including heterozygous HbE 30.2% (255/845), homozygous HbE 1.9 % (16/845), Hb Constant Spring (CS) trait 1.4% (12/845), CSEA Bart's 0.2% (2/845), and beta-thalassemia trait 0.6% (5/845). In most of the patients with diabetes, HbA1c levels determined by two different methods, inhibition immunoassay and capillary electrophoresis, gave strong positive correlation (R = 0.901, P < 0.001), except for those with homozygous HbE (N = 16) and CSEA Bart's (N = 2). In all 18 patients with homozygous HbE and CSEA Bart's, the HbA1c was undetectable by capillary electrophoresis, meaning that their estimated average glucose was undeterminable, although their HbA1c levels could be measured using an inhibition immunoassay. The discrepancy of HbA1c results obtained from two different methods is noted in patients without HbA. CONCLUSIONS: We have demonstrated the erroneous nature of HbA1c measurement in patients with hemoglobin variants, especially in those without HbA expression. Therefore, in the population with a high prevalence of hemoglobinopathies, hemoglobin typing should be considered as basic information prior to HbA1c measurement.