Enzyme immunoassay of hemoglobin A1c: analytical characteristics and clinical performance for patients with diabetes mellitus, with and without uremia.

We evaluated an enzyme immunoassay (Novo BioLabs) for determination of hemoglobin A1c and measured the contribution of carbamylated hemoglobin to the hemoglobin A1c results obtained by HPLC for patients with chronic renal failure. The reference interval, determined for blood donors, was 0.035-0.050 (i.e., 3.5% to 5% of the total hemoglobin). For non-diabetic patients with chronic renal failure, on hemodialysis, the mean proportion of hemoglobin A1c was 0.038 (SD 0.0045); for umbilical cord blood it was 0.006-0.009. For diabetic patients with normal or near-normal renal function, results by enzyme immunoassay (y) and HPLC (x) correlated well (r = 0.83): y = 0.99x - 0.005. Added fetal hemoglobin or carbamylated hemoglobin did not interfere with enzyme immunoassay, but added carbamylated hemoglobin co-chromatographed with hemoglobin A1c. Measurements for patients with chronic renal failure, with or without diabetes, by enzyme immunoassay and by HPLC, were compared with measurements for patients with diabetes mellitus but no renal failure. Carbamylated hemoglobin contributed 0.019 to the hemoglobin A1c results obtained by HPLC for patients with chronic renal failure.