Lu · Aging 2019 · cohort biomarker development and validation study · n=?

DNA methylation GrimAge strongly predicts lifespan and healthspan.

Cited 2721 times in the scientific literature.

Level 3 - non-randomized controlled study

Prospective cohort biomarker validation study across multiple clinical outcomes

PubMed 30669119 · doi:10.18632/aging.101684 · record verified 2026-08-29

What was done

The authors developed seven DNA methylation (DNAm)-based estimators of plasma proteins (including PAI-1 and GDF-15) alongside a DNAm surrogate for smoking pack-years. These surrogates were combined into a composite predictor of lifespan called DNAm GrimAge (expressed in units of years). Adjusting GrimAge for chronological age generated a measure of epigenetic age acceleration (AgeAccelGrim). The biomarker was evaluated for predicting time-to-event clinical outcomes and age-related traits using large-scale validation data from thousands of individuals.

What was found

DNAm GrimAge significantly predicted time-to-death (Cox regression P = 2.0E-75), time-to-coronary heart disease (Cox P = 6.2E-24), time-to-cancer (P = 1.3E-12), and age-at-menopause (P = 1.6E-12). It was also strongly related to computed tomography measures of fatty liver and excess visceral fat. AgeAccelGrim correlated with comorbidity count (P = 3.45E-17). Age-adjusted DNAm PAI-1 levels were associated with lifespan (P = 5.4E-28), comorbidity count (P = 7.3E-56), and type 2 diabetes (P = 2.0E-26). The abstract reports p-values but does not report effect sizes, hazard ratios, or confidence intervals.

Why it matters

GrimAge establishes a second-generation epigenetic clock framework by training DNA methylation algorithms on surrogate physiological plasma biomarkers and smoking rather than chronological age alone, substantially improving mortality and morbidity risk stratification.

Limits

The abstract does not provide exact participant counts (noted only as 'thousands of individuals'), hazard ratios, effect sizes, or confidence intervals. As an observational biomarker association study, it cannot demonstrate whether epigenetic alterations causally drive mortality or disease. Cohort demographics, ethnic diversity, and specific follow-up durations are omitted from the abstract.

Cited by