Glycemic measurements are not helpful in CVD risk prediction among non-diabetics

Several clinical guidelines recommend measurement of glycated hemoglobin (HbA1c) to guide cardiovascular risk (CVD) risk assessment. In this study, the authors examined the utility of adding HbA1c measurements to conventional risk factors in prediction of CVD among non-diabetic patients. Data from 73 prospective studies with nearly 300,000 non-diabetic patients without known CVD at enrollment was analyzed. Mean age was 58 years, 49% were women and mean HbA1c was 5.4%. Over a median follow-up duration of nearly 10 years, there were 20,840 fatal and nonfatal incident CVD events. Addition of HbA1c levels to risk prediction models with conventional CVD risk factors of age, sex, smoking, blood pressure, HDL and total cholesterol led to small changes in C-index (0.7434 to 0.7452; change of 0.0018, 95% CI 0.0003-0.0033) and net reclassification 0.42 (-0.63 to 1.48). Use of other glycemic measurements such as fasting glucose, random glucose and postload glucose were no better for CVD risk prediction than HbA1c.


Contrary to existing guidelines, addition of HbA1c levels to traditional CVD risk factors did not lead to meaningful improvement in CVD risk prediction among non-diabetics in this large, multicenter prospective study.

  • Emerging Risk Factors Collaboration, ┬áDi Angelantonio E, Gao P, et al. Glycated hemoglobin measurement and prediction of cardiovascular disease. JAMA. Mar 26 2014;311(12):1225-1233

Summarized by Steven M. Bradley and Supriya Shore