Associations of Epigenetic Age Acceleration With CVD Risks Across the Lifespan: The Bogalusa Heart Study

Abstract

Although epigenetic age acceleration (EAA) might serve as a molecular signature of childhood cardiovascular disease (CVD) risk factors and further promote midlife subclinical CVD, few studies have comprehensively examined these life course associations. This study sought to test whether childhood CVD risk factors predict EAA in adulthood and whether EAA mediates the association between childhood CVD risks and midlife subclinical disease. Among 1,580 Bogalusa Heart Study participants, we estimated extrinsic EAA, intrinsic EAA, PhenoAge acceleration (PhenoAgeAccel), and GrimAge acceleration (GrimAgeAccel) during adulthood. We tested prospective associations of longitudinal childhood body mass index (BMI), blood pressure, lipids, and glucose with EAAs using linear mixed effects models. After confirming EAAs with midlife carotid intima-media thickness and carotid plaque, structural equation models examined mediating effects of EAAs on associations of childhood CVD risk factors with subclinical CVD measures. After stringent multiple testing corrections, each SD increase in childhood BMI was significantly associated with 0.6-, 0.9-, and 0.5-year increases in extrinsic EAA, PhenoAgeAccel, and GrimAgeAccel, respectively (P < 0.001 for all 3 associations). Likewise, each SD increase in childhood log-triglycerides was associated with 0.5- and 0.4-year increases in PhenoAgeAccel and GrimAgeAccel (P < 0.001 for both), respectively, whereas each SD increase in childhood high-density lipoprotein cholesterol was associated with a 0.3-year decrease in GrimAgeAccel (P = 0.002). Our findings indicate that PhenoAgeAccel mediates an estimated 27.4% of the association between childhood log-triglycerides and midlife carotid intima-media thickness (P = 0.022). Our data demonstrate that early life CVD risk factors may accelerate biological aging and promote subclinical atherosclerosis.

Keywords: biological aging; cardiovascular disease risk factors; epigenetic age acceleration; life span; subclinical atherosclerosis.

Graphical abstract

Figure 1

Study Design In the BHS (Bogalusa Heart Study), 1,580 participants were included in the analyses to detect prospective association between childhood cardiovascular risk factors (CVRFs) and 4 forms of midlife epigenetic age accelerations (EAAs). The childhood CVRFs were aggregated using area under curve (AUC) method. Temporal relationships between factors with significant associations were further investigated among 688 participants using the cross-lagged panel analysis with simultaneously measured CVRFs and EAAs at the 2004-2006, 2008-2010, and 2013-2016 visits. A total of 1,485 participants were used to access the cross-sectional association between carotid intima media thickness (cIMT) or presence of carotid plaque and EAA. Mediation effects of identified EAA measurements were then evaluated among 531 participants.

Figure 2

Illustration of Cross-lagged Panel Analysis Model and Mediation Analysis Model (A) Cross-lagged panel analysis model of body mass index (BMI) and DNA methylation age acceleration: ρ₁ and ρ₂ are 2 cross-lagged coefficients; r₁ and r₂ are correlations between 2 instances of the same variable; and r₃ and r₄ are 2 synchronous correlations. The detailed result is shown in Table 2. (B) Mediation analyses of EAA on repeatedly measured childhood BMI with midlife cIMT. Mediator EAAs were measured at the 2008-2010 visit and cIMT was measured at the 2013-2016 visit. c is total effect; c′ is direct effect; and a × b is indirect effect. The mediation effect of EAA is calculated as (a × b / c) × 100%. The detailed result is shown in Figure 4 and Supplemental Table 6. Abbreviations as in Figure 1.

Figure 3

Associations of Childhood CVD Risk Factors With Adulthood EAA Model 1 adjusts for age, sex, and race. Model 2 adjusts for all variables in model 1 + smoking and drinking. Childhood cardiovascular disease (CVD) risk factors were summarized as an area under the curve estimate for each participant. Beta coefficients represent the change in EAA per SD increase in the CVD risk factor measure. The P values smaller than the significant threshold of 7.1 × 10⁻³ after Bonferroni correction are boldface. DBP = diastolic blood pressure; EEAA = extrinsic epigenetic age acceleration; GrimAgeAccel = GrimAge acceleration; HDL-C = high-density lipoprotein cholesterol; IEAA = intrinsic epigenetic age acceleration; LDL-C = low-density lipoprotein cholesterol; PhenoAgeAccel = PhenoAge acceleration; SBP = systolic blood pressure; TG = triglyceride; other abbreviations as in Figures 1 and 2.

Figure 4

Results of Mediation Analyses Mediation effects of EAAs on the (A) BMI-cIMT, (B) triglyceride-cIMT, and (C) HDL-C-cIMT associations in childhood and adulthood, respectively, are shown. The examined EAA measurements were selected based on our prospective and temporal analyses. Total effect, direct effect, and indirect effect are indicated as c, c′, and a × b, respectively. Abbreviations as in Figure 1, Figure 2, Figure 3.