Determination of glycosylated hemoglobin by affinity chromatography: comparison with colorimetric and ion-exchange methods, and effects of common interferences.

An affinity-chromatographic method for determination of glycosylated hemoglobin (Anal. Lett. 14: 649-661, 1981) is compared with the thiobarbituric acid colorimetric (I) (Clin. Chem. 27: 669-672, 1981) and the ion-exchange liquid-chromatographic (II) (Diabetes 29: 623-628, 1980) methods. A correlation of 0.98 was obtained for the affinity method vs II and 0.97 for affinity vs I (n = 51). The within-run CV was 1.9% for specimens from non-diabetic individuals and 1.0% for those from diabetics. The respective between-run CVs were 3.4% and 2.4%. Failure to remove "labile" glucose adducts by 5-h incubation of erythrocytes in isotonic saline (37 degrees C) contributed an average error of 13.1% for II, 5.4% for I, and 1.6% for the affinity method. Affinity chromatography gave a decrease of 0.1-0.2% glycosylated hemoglobin for each 1.0 degree C temperature increase between 18 and 27 degrees C. Varying the pH of the wash buffer used in the affinity procedure from 7.75 to 8.25 (pH 8.0 optimum) produced at net change of 0.5% in glycosylated hemoglobin with one diabetic specimen. Using the affinity method, we determined the reference interval for glycosylated hemoglobin in 124 apparently healthy individuals to be 5.3 to 7.5% (mean 6.36%, SD 0.55%). Rechromatography by II and isoelectric focusing analysis of the fractions obtained by the affinity separation revealed a substantial population of glycosylated hemoglobins not measured by II. The affinity method offers a rapid, simple, precise, and accurate alternative to methods currently in use and gives substantial freedom from many common interferences.