Comparison Between Abbott IMx and BioRad HPLC Methods for Determination of Glycosylated Hemoglobin A1c Levels in Subjects with Normal and Atypical Hemoglobins.

M. Malone, C. Dukes, T.McKaveney, S. Mehta, M. Virji and V. Warty (Department of Pathology, Univ. of Pittsburgh School of Medicine, Pittsburgh, PA 15213)

Hemoglobin A1c (HbA1c) levels reflect overall blood glucose concentration during the previous two to three months. However, hemoglobin variants can confound the measurement of this glycated hemoglobin. We compared HbA1c values from diabetic and non-diabetic subjects with normal hemoglobin and from diabetic subjects with hemoglobin AS and AC variants. Two assays using disparate methodologies, the Abbott IMx glycated hemoglobin assay and our reference, the BioRad DIAMAT assay, were used. The Abbott assay uses MEIA: ion capture to measure all glycosylated hemoglobins. HbA1c component is then calculated using a mathematical equation. It's linearity range is 4.7 to 12.7 with an interrun CV up to 6.6%. The DIAMAT uses an ion-exchange HPLC which produces individual peaks for different glycated hemoglobins including A1c. Its linearity range is 5.3 to 18.3 with an interrun CV up to 2.3%. The correlation between the two assays is as follows. In subjects with normal as well as abnormal hemoglobins: IMx=0.91DIAMAT +0.27, r2 = 0.939 (n = 90); in subjects with normal hemoglobin: IMx = 0.89 DIAMAT + 0.36, r2 = 0.956 (n = 70); and in subjects with atypical hemoglobin variants: IMx = 1.06 DIAMAT - 0.54, r2 = 0.900 (n = 20). The disparity in calculated HbA1c in subjects with normal and atypical hemoglobins in the IMx diverges positively from the DIAMAT value above 5.5% of the glycated hemoglobin. Within the clinically important range of 5-10% HbA1c, this IMx value rises to a maximum bias of 0.7% above the value obtained by BioRad. In summary, although the two methods show good correlation for these hemoglobin variants, there could be uncertainty in clinically assessing fair versus poor glucose control in patients with atypical hemoglobins by the Abbott IMx when compared with BioRad DIAMAT method.


Maintenance of normoglycemic blood glucose concentations in patients with diabetes mellitus is believed to reduce the development of diabetic complications. The presence of glucose in the circulation results in a concentration dependent nonreversible covalent glycosylation of hemoglobin, forming the hemoglobin subfractions A1a, A1b and A1c. Given the 120 day life span of the average human red blood cell, it is therefore possible to assess overall glucose levels in patients over the past two to three months by measuring the concentration of this glycosylated hemoglobin fraction. Numerous detection modalities are available commercially for measuring this fraction, including high performance liquid chromatography, affinity chromatgraphy, enzyme immunoassay and electrophoresis. Unfortunately, analytical error is introduced in patients with atypical hemoglobins, such as hemoglobin AS and AC (which represent 9% and 3% of the American black population, respectively). In our study, we compared the HbA1c values of patients with both normal and variant hemoglobins (AS and AC) run on the BioRad DIAMAT versus the Abbott IMx instruments.

Methods and Procedures

Both glycosylated hemoglobin determinations use whole blood as the specimen. The BioRad DIAMAT instrument is based on ion-exchange high performance liquid chromatography, and includes a thirty minute 37oC incubation step to prevent interference by the preglycated Schiff base. Measurement is done at two wavelengths, 415 nm and 690 nm. The recorder produces a series of peaks corresponding to the various hemoglobin fractions, the area under these peaks corresponding to the concentration in the sample. Notably, patients with variant hemoglobins often produce additional peaks on this instrument, suggesting a previously unknown hemoglobinopathy. The Abbott IMx is based on MEIA:ion capture, using a polyanion-boronate reagent, which covalently binds to the carbohydrate moiety of the glycosylated hemoglobin. This polyanion-hemoglobin complex is then "captured" in a plastic, positively charged matrix, and hemoglobin concentration is determined by its quenching the fluorescence of 4-methylumbelliferone. A subsequent measurement of the hemoglobin concentration in the entire sample is made, and the ratio of glycosylated to total hemoglobin calculated. The fraction of hemoglobin A1c is based on the mathematical formula:
%A1c= %GHb + 1.76/1 .49

Clinical Protocol

Ninety patient samples were run in parallel on the two instruments with HbA1c amounts ranging from 3.7% to 15.1%. Seventy of these samples had normal hemoglobin, and represented a mixture of diabetic and non-diabetic patients. The remaining twenty patients, all diabetics, had the variant hemoglobins AS (n = 14) and AC (n = 6). The DIAMAT required 5 uL of sample, and the IMx 150 uL. Samples were collected in EDTA anticoagulant, and stored at 4oC for not greater than seven days. Results were plotted on a two-dimensional curve, with regression analyses performed to compare the DIAMAT versus IMx results on all samples, normal hemoglobins and atypical hemoglobins.

Results and Discussion


Who's who -- wwwadmin