HbA1c may not be a sensitive determinant of diabetic status in the elderly

Abstract Objective American Diabetes Association (ADA) has recently recommended the use of glycated haemoglobin (HbA1c) to diagnose diabetes mellitus. We aim to determine if indeed this recommendation applies to the population in Singapore and whether it varies with age. Method This is a cross sectional study of 90 patients without previous history of diabetes who underwent screening and had both oral glucose tolerance test (OGTT) and HbA1c done at the same time. These patients were stratified into 4 age groups. Result We found that HbA1c of 6.2% is the best cut-off to diagnose diabetes using ROC curve analysis. At the specified HbA1c, the area under ROC curve (AUROC) reduces as age group increases suggesting that sensitivity and specificity of HbA1c as diagnostic marker reduces as age increases. Conclusion HbA1c has a low sensitivity to diagnose diabetes in older Asian subjects and caution is required when using HbA1c in isolation. This raises the possibility that a different cut-off value for different age groups may be more appropriate.