Abstract
Psychosocial adversity over the life course may impact the aging process. However, life-course models have yet to fully explain the biological embedding of psychosocial adversity in aging. A subsample of participants from the Hispanic Community Health Study/Study of Latinos with DNA methylation (DNAm) data (N=970) was used to evaluate the effect of adversity on biological aging and the most compatible life-course model. Epigenetic age was estimated from GrimAge and DunedinPACE. We modified a current Bayesian life-course model to estimate the effect of adversity from childhood to adulthood on epigenetic age acceleration and the weights contributed by childhood (W (childhood) ) and adulthood (W (adulthood) ), which sum to one. The model was also used to evaluate the compatibility with the sensitive period (W (childhood) ≠ W (adulthood) ) and accumulation models (W (childhood) ≈ W (adulthood) ). Causal mediation analysis assessed the pathway model by estimating the indirect effect of childhood adversity through adulthood adversity. Per-unit increase in adversity was associated with 0.91 years (95% credible interval [CrI]: 0.28, 1.53; W (adulthood) = 82%, 95% CrI: 36%, 99%) increased GrimAge acceleration (AgeAccelGrim) and 0.013 years/calendar year (95% CrI: -0.005, 0.032; W (childhood) = 49%, 95% CrI: 3%, 97%; W (adulthood) = 51%, 95% CrI: 3%, 97%) increased DunedinPACE. A pronounced indirect effect of childhood adversity was found in AgeAccelGrim (0.23 years, 95% CI: 0.09, 0.37) but minimal in DunedinPACE (0.003 years/calendar year, 95% CI: -0.001, 0.006). Psychosocial adversity from childhood to adulthood may affect biological aging, with distinct life-course models explaining its effects on different aging markers.