Comparison of hemoglobin A1c and hemoglobin A1 in diabetic patients

Hemoglobin A1 (HbA1) and hemoglobin A1c (HbA1c) values were compared in 934 samples from 686 subjects with diabetes mellitus. Linear regression analysis revealed the two parameters to be highly related with HbA1 = 1.18 HbA1c + 1.67, and a correlation coefficient of 0.97. Samples with HbA1 greater than predicted from the measured HbA1c had been stored significantly longer (∼1 day) than other samples, suggesting that sample storage may slightly alter the relationship between these parameters. Serial HbA1 and HbA1c values in all 9 patients for whom data was available from at least 5 time points changed in parallel regardless of the time interval between sampling or the magnitude of the interval change of HbA1 and HbA1c. These data indicate that the potentially simpler measurement of HbA1 gives results equivalent to HbA1c for clinical applications. There is considerable interest in the measurement of HbA1 and HbA1c as tools for monitoring patients with diabetes mellitus since they appear to correlate with the time-averaged plasma glucose concentration over the preceeding 3 to 6 weeks (2-18). Most workers have studied HbA1c (4-9) but the techniques required to measure this species (chomatography, isoelectric focusing) tend to be time-consuming and are not well suited to the routine clinical laboratory. There have been some reports on relatively small numbers of patients in which HbA1c has been measured along with HbA1_a+b or HbA1. Most of these reports indicate that HbA1_{a + b} is elevated in diabetics but to a lesser extent than HbA1c (4, 13-15) although one report found no difference in HbA1_{a + b} values between diabetics and controls (16). Hemoglobin A1_{a + b} has been found to correlate significantly with HbA1c with correlation coefficients of 0.73 (13), 0.49 (14), and 0.82 (17). Moreover, a correlation coefficient of 0.996 for HbA1c versus HbA1 has been reported for 38 diabetic and 28 control subjects (18). The studies reported here were performed to confirm and extend the above findings in a large population of diabetic patients in order to determine whether values for the potentially more easily measured HbA1 are equivalent to HbA1 values for clinical use.