Background:
DNA methylation (DNAm)-based measures of aging, termed epigenetic clocks (EC), are associated with aging-related outcomes including cardiovascular disease (CVD) and all-cause mortality. Associations of ECs with heart failure (HF) are unclear. We tested whether ECs were positively associated with risk of incident HF in the Atherosclerosis Risk in Communities (ARIC) study and evaluated whether adding ECs to Pooled Cohort Equation (PCE) variables improved risk prediction.
Methods:
We measured DNAm in peripheral blood leukocytes in 2,263 African American (mean age=56.3 years) and 925 European American (mean age=59.5 years) participants using the Illumina HM450K and calculated 7 ECs: Horvath, Hannum, extrinsic (EEAA) and intrinsic (IEAA) epigenetic age acceleration, Hannum, PhenoAge, and GrimAge. HF was ascertained by
ICD-
9 code 428 and adjudication by an expert panel. We carried out race stratified proportional hazards regression to test associations ECs with incident HF, adjusting for PCE variables: age, sex, smoking, total cholesterol, HDL, systolic blood pressure (SBP), antihypertensive medication use, and diabetes. We calculated area under the curve (AUC) and integrated discrimination index (IDI) to evaluate improvement in risk prediction when adding the ECs to PCE variables.
Results:
The number of incident HF events and mean follow-up time in African Americans and European Americans were 640 (189 in the first 10 years) and 19.3 years, and 191 and 21.7 years, respectively. All 7 ECs were positively associated with HF in both African Americans and European Americans. In African Americans with follow-up restricted to the first 10 years, the HR for a one SD increment in GrimAge (5.64 years) was 1.57 (95% CI=1.31, 1.88), comparable to that for a one-SD (5.82 years) increment in age (HR=1.58, 95% CI=1.36, 1.83) and greater than that for a one-SD (20.2mmHg) increment in SBP (HR=1.33, 95% CI=1.18, 1.51). In European Americans across the entire follow-up period, the HR for a one-SD increment in GrimAge (6.13 years) was 1.22 (95% CI=1.06, 1.41), smaller than that for a one-SD (5.50 years) increment in age (HR=1.93, 95% CI=1.63, 2.29) and larger than that for a one-SD (17.9 mmHG) increment in SBP (HR=1.13, 95% CI=0.98, 1.30). In African Americans with follow-up restricted to the first 10 years, adding GrimAge to PCE variables increased AUC by 0.019 (95% CI=0.003, 0.035) and the IDI was 0.010 (95% CI=0.002, 0.019). In European Americans, adding GrimAge did not change AUC appreciably (0.004, 95% CI=-0.006, 0.014) and the IDI was 0.002 (95% CI=0.000, 0.005).
Conclusion:
ECs are positively associated with HF in African American and European American participants independent of traditional CVD risk factors. GrimAge modestly improved heart failure risk prediction in African Americans. HF-specific DNAm-based measures should be developed and evaluated for improvement in risk prediction.
Cardiovascular Risk Factor Targets and Cardiovascular Disease Event Risk in Diabetes: A Pooling Project of the Atherosclerosis Risk in Communities Study, Multi-Ethnic Study of Atherosclerosis, and Jackson Heart Study