Abstract 12975: HbA1c, Glycemia, and Future Cardiovascular Events: The Strong Heart Study
Introduction: Fasting plasma glucose (FPG) has traditionally been used to diagnose diabetes. HbA1c≥6.5% is now accepted as an alternative criterion. The comparative value of using HbA1c vs. FPG to predict cardiovascular disease (CVD) risk has not been examined in populations with widespread obesity and diabetes.
Methods: Our population included 4549 American Indians, ages 45–74 yrs, from the baseline Strong Heart Study exam (1989–91). Data from 3850 (60% women) who were free of CVD and had baseline measures of FPG and HbA1c were analyzed; 1386 had known diabetes. CVD events were ascertained over a median of 15 yrs. Cox proportional hazards models, adjusted for known CVD risk factors, were used to test the relationship of glycemia, as measured by HbA1c or FPG, with incident coronary heart disease (CHD) or total CVD.
Results: An HbA1c > 6.5% had 44.3% sensitivity and 98.9% specificity to identify participants who had FPG ≥ 126 mg/dL. Increases in HbA1c were associated with higher baseline prevalence of CVD risk factors; those with known diabetes had the worst CVD risk factor profiles. For HbA1c <5%, [5, 5.5%), [5.5, 6%), [6, 6.5%), ≥6.5%, or those with known diabetes, the multivariate-adjusted HR for CHD was significant only for those with HbA1c > 6.5% (HR = 1.49; 95% CI = 1.01, 2.19) or with known diabetes (HR = 2.99; 95% CI = 2.35, 3.80), using HbA1c < 5% as the referent group. Similarly, the adjusted HRs for total CVD were significant only for those with HbA1c > 6.5% or with known diabetes (1.58, 95% CI = 1.15, 2.16 and 2.73, 95% CI = 2.23, 3.34, respectively). Similar results were observed for FPG (i.e., HRs were significant only for FPG >126 or known diabetes). In participants with diabetes, neither HbA1c nor FPG as a continuous variable was significantly associated with incident CHD or total CVD in any of the adjusted models.
Conclusions: In this population with high prevalence of diabetes, only those with newly diagnosed diabetes (as measured by HbA1c or FPG) or known diabetes had higher incidence of CVD. In persons with diabetes, neither HbA1c nor FPG was an independent predictor of CVD. While using HbA1c instead of FPG to diagnose diabetes is more convenient, neither test adds to conventional CVD risk factors in predicting CHD or total CVD outcomes.
- © 2010 by American Heart Association, Inc.