Increased carbamylated hemoglobin formed in erythrocytes during uremia may
interfere with HbA(1c) assays, but few studies compared directly both param
eters. We measured carbamylated hemoglobin by HPLC in 45 non-diabetic uremi
c patients (16 with acute and two with chronic renal failure, 27 with trans
plant recipients) as 57.8 +/- 22.3 mu g carbamylvaline /g Hb (mean +/- stan
dard deviation) vs. 31.6 +/- 5.1 in 15 controls (+83%, p < 0.001). In these
samples, HbA(1c) was evaluated by three ion-exchange HPLC methods, 1: Diam
at (BioRad), 2: A1c2.2 (Tosoh) and 3: HA8140 (Menarini), and one immunoassa
y method (Tinaquant II Roche). Whichever the method, mean HbA(1c) values ob
tained increased in patients with high (> 60 mu g carbamylvaline/g Hb) vs.
low (< 45) carbamylated hemoglobin values (+0.08 to 0.25% of total Hb), but
differences were not significant. Minor peaks on the chromatograms were ho
wever increased in parallel to carbamylated hemoglobin. HbA(1c) values over
6% were found in 4, 1, 2 and 0 samples, with HPLC 1,2, 3 and immunoassay,
respectively. Fructosamine values were not significantly altered. Our resul
ts show that Hb adducts, whether due to carbamylation or to other chemical
reactions, interfere to a variable extent with different HbA(1c) assay meth
ods, and confirm that HbA(1c) values should be interpreted with caution in
uremic patients.