Evaluation of error levels in hemoglobin A1c and glycated albumin in type 2 diabetic patients due to inter-individual variability.Diabetes Res Clin Pract. 2010 Aug; 89(2):115-20.DR
To evaluate error levels in hemoglobin A1c (A1C) and glycated albumin (GA) in type 2 diabetic patients due to inter-individual variability.
Type 2 diabetic patients with stable glycemic control and without complications affecting either A1C or GA were enrolled (n=154; age 68.4+/-9.9 years). Blood examination was performed 1-4h after breakfast or lunch every 2-3 months on > or =3 occasions. A1C data were changed to IFCC values for analysis.
A1C and GA correlated significantly with postprandial plasma glucose. The correlation coefficient between A1C and GA was 0.728 (p<0.001) when calculated using raw data and 0.747 (p<0.001) when calculated using averaged data for each patient. The ratio R of GA to A1C was 3.88+/-0.50 for raw data and 3.88+/-0.47 for averaged data, indicating coefficients of variation of R (CV(R)) of 12.9% and 12.1%, respectively. Multiple regression analysis reduced CV(R) to 11.2%. After dividing CV(R)(2) into CV(A1C)(2) and CV(GA)(2), CV(A1C) and CV(GA) were calculated as 9.1% for raw data and 8.6% for averaged data, and were reduced to 7.9% after multiple regression analysis.
Error levels in A1C and GA reach 7.9-9.1%, suggesting the existence of maximal 18% errors in A1C and GA levels.