Fig. 2
Cumulative lifetime stress is associated with epigenetic age acceleration in a highly traumatized human cohort derived from the Grady Trauma Project. Epigenetic age acceleration (Δ-age) was calculated by subtracting chronological age from DNA methylation predicted age. Δ-age was regressed on cumulative lifetime stress (Life Stress) after adjusting for covariates (fitted stress measures are shown). a Life Stress was positively associated with epigenetic age acceleration (β = 0.18, SE = 0.08, P = 1.8 × 10−2), and this association remained significant after further controlling for lifestyle parameters, including body mass index, smoking, alcohol, cocaine, marijuana, and heroin use (β = 0.31, SE = 0.11, P = 7.4 × 10-3), as well depressive symptomatology, psychiatric treatments, and genome-wide SNP-based principal components (β = 0.28, SE = 0.13, P = 2.7 × 10−2). Statistically significant association was found for Personal Life Stress (β = 0. 26, SE = 0.10, P = 8.7 × 10−3) (b), whereas the effect of Network Life Stress was not significant (P = 1.1 × 10−1) (c). Age stratification by a median split showed that the effect of Personal Life Stress on Δ-age was stronger in older (β = 0.33, SE = 0.17, P = 5.3 × 10−2) (d), as compared to younger participants (β = 0.15, SE = 0.14, P = 2.8 × 10−1) (e). Stratification of the effect of cumulative life stress on epigenetic age acceleration based on the presence or not of moderate to severe physical or sexual child abuse showed that Life Stress was positively associated with Δ-age in participants with no or mild physical and sexual child abuse (β = 0.34, SE = 0.11, P = 2.5 × 10−3, n = 212) but not in those with moderate to extreme child abuse (P = 3.9 × 10−1, n = 174) (f)