A comparison of glycated albumin and glycosylated hemoglobin for the screening of diabetes mellitus in Taiwan.

Objective Glycated albumin (GA) values reflect an average plasma glucose level over approximately 2–4 weeks, and the assay is stable and can be run on serum or plasma. The aim of this study was to determine the universality and the clinical utility of GA in screening for diabetes mellitus. Methods Subjects consisted of 2192 male and female residents in Yi-lan County, Northern Taiwan (mean age 60.1 years), of whom 54.2% (n = 1188) had previously been diagnosed and treated for diabetes. Fasting blood samples were obtained to measure HbA1c, plasma glucose, serum GA, insulin, and measures of kidney and liver function. The reference values for these parameters were determined. Data from patients with diabetes and non-diabetic controls were also compared. Results Mean GA values were 13.8% in controls and 18.1% in diabetic subjects (31.2% higher, p < 0.0001), while mean HbA1c values were 5.6% in controls and 7.2% in diabetic subjects (29.2% higher, p < 0.0001). The 95th percentile values for GA and HbA1c in controls were 16.1% and 6.2%, respectively. Our suggested GA and HbA1c cut-points for prediabetes at the 75th percentile of the normal population would be 14.6% and 5.8%, respectively. For both parameters, values greater than these cut-points provided a reasonable degree of specificity and sensitivity for risk of having diabetes, while a GA value of 16.5% corresponds to an HbA1c level of 6.5%, diagnostic of diabetes. Conclusion These data indicate that GA values can be used as a surrogate parameter for HbA1c in screening for prediabetes and diabetes mellitus. [ABSTRACT FROM AUTHOR]