BACKGROUND: Patients with hemoglobin (Hb) variants may produce false HbA1c measurement. This study aimed to detect the common Hb variants in southern Taiwan and to evaluate their effect on the determination of HbA1c. METHODS: A total of 1,434 samples collected for HbA1c measurement at Kaohsiung Veterans General Hospital in southern Taiwan in March 2008 were submitted for Hb variant analysis by Primus CLC-385. HbA1c measurements were obtained using ion-exchange high-performance liquid chromatography (HPLC) (Tosoh HLC-723 G7) for routine analysis. Patients identified with Hb variants were recalled for boronate-affinity HPLC analysis. The values of estimated average glucose (eAG) were converted from HbA1c. Values of eAG-FPG, calculated by eAG minus fasting plasma glucose (FPG), were compared to estimate the accuracy of HbA1c measurement in patients with Hb variants. RESULTS: Among the 1,434 patients, the mean standard deviation of FPG was 162.8 +/- 60.5 mg/dL, HbA1c was 8.28 +/- 1.97%, and eAG was 190.9 +/- 56.6 mg/dL. Five Hb variants were detected in 11 patients, the incidence being 0.76%. Hb J was identified in 4 patients, Hb G in 2 patients, Hb E in 1 patient, Hb owari in 3 patients, and high fetal hemoglobin (HbF) in 1 patient. Abnormal HPLC chromatograms were seen among the patients with Hb J, E, G and HbF, but not in the patients with Hb owari. In patients with Hb variants, FPG was 149.5 +/- 39.9 mg/dL, HbA1c was 7.29 +/- 2.01%, and eAG was 162.5 +/- 57.7 mg/dL. Lower values of eAG-FPG may have occurred in the patients with Hb J and E, and in those with high HbF. On scattergrams of the relationship between HbA1c and FPG, the plots of Hb J, E and high HbF lay below the regression line of non-Hb variants. Inconsistent Hb values between both methods were only observed among some samples of patients with Hb variants. CONCLUSION: The existence of Hb variants may result in false HbA1c measurement. The possible presence of spuriously low HbA1c levels or abnormal HPLC chromatograms by using ion-exchange methods should be kept in mind.
BACKGROUND:Patients with hemoglobin (Hb) variants may produce false HbA1c measurement. This study aimed to detect the common Hb variants in southern Taiwan and to evaluate their effect on the determination of HbA1c. METHODS: A total of 1,434 samples collected for HbA1c measurement at Kaohsiung Veterans General Hospital in southern Taiwan in March 2008 were submitted for Hb variant analysis by Primus CLC-385. HbA1c measurements were obtained using ion-exchange high-performance liquid chromatography (HPLC) (Tosoh HLC-723 G7) for routine analysis. Patients identified with Hb variants were recalled for boronate-affinity HPLC analysis. The values of estimated average glucose (eAG) were converted from HbA1c. Values of eAG-FPG, calculated by eAG minus fasting plasma glucose (FPG), were compared to estimate the accuracy of HbA1c measurement in patients with Hb variants. RESULTS: Among the 1,434 patients, the mean standard deviation of FPG was 162.8 +/- 60.5 mg/dL, HbA1c was 8.28 +/- 1.97%, and eAG was 190.9 +/- 56.6 mg/dL. Five Hb variants were detected in 11 patients, the incidence being 0.76%. Hb J was identified in 4 patients, Hb G in 2 patients, Hb E in 1 patient, Hb owari in 3 patients, and high fetal hemoglobin (HbF) in 1 patient. Abnormal HPLC chromatograms were seen among the patients with Hb J, E, G and HbF, but not in the patients with Hb owari. In patients with Hb variants, FPG was 149.5 +/- 39.9 mg/dL, HbA1c was 7.29 +/- 2.01%, and eAG was 162.5 +/- 57.7 mg/dL. Lower values of eAG-FPG may have occurred in the patients with Hb J and E, and in those with high HbF. On scattergrams of the relationship between HbA1c and FPG, the plots of Hb J, E and high HbF lay below the regression line of non-Hb variants. Inconsistent Hb values between both methods were only observed among some samples of patients with Hb variants. CONCLUSION: The existence of Hb variants may result in false HbA1c measurement. The possible presence of spuriously low HbA1c levels or abnormal HPLC chromatograms by using ion-exchange methods should be kept in mind.