Associations between HbA1c-derived estimated average glucose and fasting plasma glucose in patients with normal and abnormal hemoglobin patterns.

Fasting plasma glucose (FPG) and HbA1c are well-known tests for monitoring short and long-term glycemic control, respectively. Estimated average glucose (eAG) is derived from the HbA1c for expression in the same units as FPG. The present study aimed to evaluate the association between eAG and FPG values. This retrospective study was performed on 1285 patients who measured the FPG and HbA1c (capillary electrophoresis method) values on the same day. They were initially divided into 3 hemoglobin-pattern groups, normal, HbE, and high HbA2, and then these 3 groups were further subdivided into whole group, subgroup 1 (HbA1c ≤ 7%) and subgroup 2 (HbA1c >7%) for analytical comparisons. The eAG values were calculated using Nathan's equation. The differences, agreements, and correlations between eAG and FPG were evaluated. Good agreements and strong positive correlations between eAG and FPG values were observed in all 3 whole groups. In subgroup analysis, the degrees of agreement and correlation depended on the level of glycemic control, as all associations became stronger with better glycemic control. Additionally, the HbA1c levels and different eAG-FPG values between the normal and other groups were not significantly different in either whole group or subgroup comparisons. Our study found good agreements and strong positive correlations between eAG and FPG in all groups, indicating that high HbA2 or HbE did not affect the HbA1c levels or thus the associations between eAG and FPG. Reporting the eAG together with the HbA1c value should help improve the understanding of glycemic status in patients, leading to improved blood glucose control.

Strong correlation and high comparability of capillary electrophoresis and three different methods for HbA1c measurement in a population without hemoglobinopathy.

The present study aimed to investigate the comparability of capillary electrophoresis (CE) and three different methods for HbA1c measurement. The 270 whole blood samples with normal hemoglobin profiles were analyzed for HbA1c values by the Roche turbidimetric inhibition immunoassay (TINIA), the Mindray enzymatic assay (EA), the Arkray high-performance liquid chromatography (HPLC) in comparison with the Sebia CE. The within-laboratory coefficient of variation of all four methods was within the accepted goal (<2%), showing good performance of all of these methods. Pairwise comparisons of HbA1c values obtained by CE and other methods were determined in both total group and subgroups (HbA1c levels of <6.5%, 6.5-8% and >8%). Mean differences of HbA1c values in all groups were very small in which the mean HbA1c value measured by EA was lower while those by TINIA and HPLC were higher than that of CE. The majority of different values were within the limits of agreement in Bland-Altman analysis, indicating good agreement between CE and the others. Less than 5% of percentage differences were out of the total allowable error limit in all groups, showing that differences of HbA1c values between CE and other methods were not clinically significant. Pairwise comparisons of HbA1c values of CE and the others in Passing-Bablok regression and Spearman rank correlation studies displayed high concordance and strong correlation in all groups. In conclusion, the present study showed strong correlation, high comparability and consistent results for HbA1c measurement between capillary electrophoresis and the other three different methods.