Effect of Hemoglobin Variants on Routine Glycohemoglobin Measurements Assessed by a Mass Spectrometric Method

Comparative analyses of glycohemoglobin (HbA1c) in samples containing hemoglobin (Hb) variants have shown that different test systems may give discrepant results (1)(2)(3). HPLC methods for such samples generally underestimate the true HbA1c value, although a few variants give a positive error for HbA1c. Immunoassays may also underestimate the values if there is a change in an epitope that contains glucose and N-terminal amino acids of the Hb β chains. Therefore, more information needs to be collected on the effects of various Hb variants on specific HbA1c test systems (1). The reference method proposed by Kobold et al. (4) for measuring HbA1c is based on electrospray ionization mass spectrometry (ESI/MS) determination of the N-terminal residues of the Hb β chains, which are released by enzymatic cleavage of the intact Hb molecule with endoproteinase Glu-C. This method gives accurate results for the percentage of glycated Hb at the N terminus of the β chains even for samples containing Hb variants. In the present study, we compared the HbA1c values measured in samples with various Hb variants by two commercial systems (HPLC and immunoassay) and the ESI/MS method. A total of 81 samples, from nondiabetic and diabetic subjects, were analyzed, of which 45 were homozygous for HbA and 36 were heterozygous for various Hb variants (Table 1⇓ ). The structures of most Hb variants were determined by MS and by DNA analysis, primarily to elucidate the cause of the unexpected values of HbA1c measured by HPLC; one case (HbM Boston) was examined for cyanosis-like symptoms (5). For high-resolution HPLC to measure the content of the variants, we used cation-exchange column chromatography with Polycat A packing and a slow (90 min) gradient buffer change (6). Because HbM was not separated by …