[Relationship between glycated albumin (GA) and glycated hemoglobin (A1c) in 255 patients with liver diseases using cross-sectional laboratory data].

To investigate how liver disease alter the serum glycated proteins as markers of diabetic control, we studied serum GA, A1c and especially GA/A1c ratio in 255 patients having over 35IU/L in ALT(transaminase) compared with those of 829 type 2 diabetes mellitus (DM) in cross sectional manner. 255 patients with liver diseases were divided into 69 patients with biopsy proven liver cirrhosis (LC), 66 patients with chronic hepatitis(CH) and 120 patients with fatty liver(FL) diagnosed by abdominal echography. The mean GA/A1c ratio (+/-SD) was significantly higher (p<0.0001) in LC group(3.71+/-1.03) than the other groups (3.03+/-0.45 for CH, 3.05+/-0.42 for DM), while the mean GA/A1c ratio in FL group was significantly lower(2.74+/-0.31) (p<0.0001)) than that of DM groups. In LC group the GA/A1c ratio increased significantly depending upon serum albumin and/or platelet reductions. The GA/A1c ratio was significantly correlated with the other laboratory data such as serum albumin, cholinesterase, total cholesterol levels and weakly correlated with serum hemoglobin level. We also followed the serum levels of GA and A1c and the GA/A1c ratio during about 13 months (5 times blood collections) in 18 patients enrolled in this study. Resultantly the coefficient of variation of GA/A1c ratio was the smaller than the others(GA, A1c). The ROC curve of GA/A1c ratio for LC versus FL group was the most reliable between four groups and the cut-off value for LC versus FL was 2.94. Theses results suggest that GA/A1c ratio could be an useful marker for different diagnosis when facing patients with abnormal serum ALT level in a clinical setting.