Performance characteristics and clinical utility of an enzymatic method for the measurement of glycated albumin in plasma.

OBJECTIVE The measurement of plasma glycated albumin is particularly useful in the short-middle term monitoring of glycometabolic control in diabetics. The aim of this work is to evaluate a new enzymatic method for the measurement of glycated albumin in plasma, with particular attention to some selected cases and comparison with other relevant tests (fasting plasma glucose, after glucose load, fructosamine, glycated hemoglobin). DESIGN AND METHODS We have performed a multicenter study by which sample collection was performed in three different centers (Milano, Padova and Cagliari) and serum samples, frozen at -80 degrees C, were then delivered under dry ice to the centralized laboratory in Milano. Glycated plasma albumin was measured with reagents from Asahi Kasei Pharma (Lucica GA-L enzymatic assay; AKP, Tokyo, Japan) on a Modular P Roche system. Fructosamine was assessed by a Roche method and HbA(1c) (measured separately in the three centers on fresh EDTA blood) by DCCT-aligned HPLC systems. We have investigated 50 type 2 diabetics, 26 subjects with gestational diabetes, 35 subjects with thalassemia major, 10 subjects with cirrhosis, 23 patients with end-stage renal disease subjected to dialysis treatment and 32 healthy adult control subjects. RESULTS The main analytical performance characteristics of the new GA test were the following: (a) the within-assay reproducibility was between 3.0 and 3.9% (in terms of GA% CV, measured on 2 serum pools and 2 control materials at normal and pathological glycated albumin levels); (b) the between-assays reproducibility was from 2.8 to 4.1%; (c) the linearity was tested in the interval between 13 and 36% and found acceptable (r(2)=0.9932). Concerning the clinical utility of the new test, we have evaluated the relationships between GA, HbA(1c), fructosamine and fasting and post-prandial glucose in several patients, as well as the changes in the above mentioned parameters in a sub-group of type 2 diabetic patients for 18 weeks as they progressed from severe hyperglycemia (HbA(1c) >or=10.0%) toward a better glycemic control. The correlations between glycated albumin and HbA(1c) were as follows: (a) type 2 diabetics: r(2)=0.483 (good glycemic control), r(2)=0.577 (poor control); (b) diabetic patients under dialysis: r(2)=0.480; (c) liver disease: r(2)=0.186; (d) transfused non-diabetics with thalassemia: r(2)=0.004. Glycated albumin, as well as HbA(1c) and fructosamine, was of little value in the study of women with gestational diabetes, mainly because of the very limited glucose fluctuations in this particular category of subjects. In 11 type 2 diabetic patients under poor metabolic control, GA was better correlated with fasting plasma glucose then HbA(1c) (r(2)=0.555 vs. 0.291, respectively), and decreased more rapidly than HbA(1c) during intensive insulin therapy. CONCLUSIONS The experience we have acquired with the new enzymatic test demonstrates its reproducibility and robustness. We confirm that plasma glycated albumin is better related to fasting plasma glucose with respect to HbA(1c). Moreover, glycated albumin is more sensitive than HbA(1c) with regard to short-term variations of glycemic control during treatment of diabetic patients. This test is also very appropriate when the interpretation of HbA(1c) is critical.