Mortality associations with DNA methylation-based biological aging and physical functioning measures across a 20-year follow-up period

Abstract

Background Measures of biological aging range from DNA methylation (DNAm)-based estimates to measures of physical abilities. The purpose of this study was to compare DNAm- and physical functioning-based measures of biological aging in predicting mortality.

Methods We studied 63- to 76-year-old women (N = 395) from the Finnish Twin Study on Aging (FITSA). Participants’ biological age (epigenetic clocks DNAm GrimAge and DunedinPACE) was estimated using blood DNAm data. Tests of physical functioning conducted under standardized laboratory conditions included the Timed Up and Go (TUG) test and 10-m walk test. Mortality hazard ratios (HRs) were calculated per every one standard deviation (SD) increase in the predictor. Cox regression models were conducted for individuals and twin pairs, the latter controlling for underlying genetic effects. The models were adjusted for known lifestyle predictors of mortality.

Results During the follow-up period (mean 17.0 years, range 0.2–20.3), 187 participants died. In both the individual-based and pairwise analyses, GrimAge and both functional biomarkers of aging were associated with mortality independent of family relatedness, chronological age, physical activity, body mass index, smoking, education, or chronic diseases. In a model including both the DNAm-based measures and functional biomarkers of aging, GrimAge and TUG remained predictive.

Conclusions The findings suggest that DNAm GrimAge and the TUG test are strong predictors of mortality independent of each other’s and genetic influences. DNAm-based measures and functional tests capture different aspects of the aging process and thus complement each other as measures of biological aging in predicting mortality.