Abstract
BACKGROUND: Cardiovascular (CV) health-related risk factors may influence the epigenetic-based pace of biological ageing (BA). However, given that early-life lifestyle factors can have lasting effects on DNA methylation (DNAm), observed associations may reflect cumulative exposures rather than short-term changes in older adults. We investigated whether CV risk factors are associated with changes in the pace of ageing in midlife and older adults. METHODS: We analysed baseline DNA methylation data from 4848 participants across three cohorts, AGES-RS (n = 2602), CARDIA (n = 1568), and InCHIANTI (n = 678), using Illumina arrays, with two to four time points per cohort. Pace of ageing was measured using DunedinPACE (DDPACE). Cardiovascular risk factors included smoking status, DNAm-derived pack-years, physical activity (PA), body mass index (BMI), systolic and diastolic blood pressure (SBP and DBP), total cholesterol, blood fasting glucose, and their composite score (adapted Life's Simple 7 [adapted-LS7]). We conducted three analyses: (1) prospective analysis of CV risk factors (exposures) and DDPACE (outcome); (2) delta analysis of changes in DDPACE; and (3) shift analysis focusing on participants whose pace of ageing accelerated (accelerators: ≥1 SD above the mean) or decelerated (decelerators: ≤1 SD) over time. After excluding individuals with consistently average, fast, or slow ageing patterns, we conducted each analysis at the 5-year and 9(+)-year follow-ups. FINDINGS: Across cohorts, >55% were female. Mean (SD) age ranged from 40.3 (3.6) in CARDIA to 76.3 (5.2) in AGES-RS. DDPACE also varied: 0.92 (0.13) in CARDIA vs. 1.10 (0.11) in AGES-RS. Within 5-year follow-up, smoking (current and former), pack-years of smoking, BMI, and blood glucose level were longitudinally associated with faster ageing (higher DDPACE; P < 0.05 (two-sided, linear mixed modeling)) in each cohort and in the meta-analyses. PA, diet, and higher adapted-LS7 scores were linked to slower ageing. Delta analyses confirmed consistent associations in AGES-RS, CARDIA, and the meta-analysis. In the meta-analysis of the 5-year shift model, higher pack-years, BMI, SBP, and DBP increased odds of being in the "accelerator" group compared to a "decelerator" (P < 0.05 (two-sided, logistic regression)), while higher adapted-LS7 scores increased odds of being a "decelerator" (P < 0.05 (two-sided, logistic regression)). Midlife prospective analysis in AGES-RS (∼age 50) showed similar patterns. Longer-term follow-ups (9+ years) in CARDIA and InCHIANTI showed similar but less consistent findings with the 5-year follow-up. INTERPRETATION: Targeting modifiable CV risk factors, particularly smoking, physical inactivity, and poor cardiometabolic health, may help slow ageing and reduce age-related disease burden in midlife and older adults. FUNDING: National Institute on Aging, National Heart, Lung, and Blood Institute, Icelandic Heart Association, and Italian Ministry of Health.