GrimAge (DNAm GrimAge Epigenetic Clock)

GrimAge is a second-generation epigenetic clock built to predict time-to-death and time-to-disease rather than chronological age itself. Ake Lu and colleagues at UCLA trained it on DNA methylation surrogates for seven plasma proteins tied to mortality risk plus a methylation-based estimate of smoking pack-years, then regressed the whole composite on time-to-death [1]. The claim: because it targets mortality directly during training, GrimAge should outperform first-generation clocks like Horvath's at predicting who dies sooner and who develops heart disease, cancer, or early menopause.

The original 2019 validation, run on more than 7,000 Illumina array measurements pooled across cohorts, found GrimAge outperformed prior clocks for time-to-death (Cox regression p=2.0x10^-75), time-to-coronary-heart-disease (p=6.2x10^-24), time-to-cancer (p=1.3x10^-12), and age-at-menopause (p=1.6x10^-12) [1]. It also correlated with CT-measured visceral and liver fat. These are large, statistically overwhelming associations in observational cohort data, not trial-based causal proof that lowering GrimAge extends life.

How to measure it: a blood draw is processed on an Illumina methylation array (450K or EPIC), then the published Lu et al. coefficients calculate the seven DNAm protein surrogates, the smoking-pack-years estimate, and the final composite age [3]. Commercial providers, including TruDiagnostic, return GrimAge alongside Horvath, PhenoAge, and DunedinPACE as one panel [5].

How to intervene on it: no drug has regulatory approval to lower GrimAge. The uncontrolled TRIIM pilot (n=9, growth hormone plus DHEA and metformin) reported a roughly 2.1-year GrimAge increase that did not reverse after treatment stopped, unlike the other three clocks tested in the same trial, illustrating that different clocks can move in different directions under the same intervention [2]. Smoking cessation directly lowers the smoking-pack-years component, giving it one of the more mechanistically direct levers among all clocks. Exercise and diet quality show smaller, inconsistent associations in cohort data.

Critics, including some of the field's own methodologists, note GrimAge partly encodes smoking history and protein levels already known to predict mortality, so part of its predictive power is not mysterious epigenetic insight but a sophisticated repackaging of known risk factors into a methylation signature [4][6]. It also has not been validated in a randomized trial as a causal target: a drug that lowers GrimAge might not lower actual mortality risk.

The plain takeaway: GrimAge is currently the strongest cohort-validated epigenetic mortality predictor among widely used clocks, useful as a risk-stratification tool, but changing it with a supplement or drug is not the same as proven life extension, and no clinical trial has demonstrated that manipulating GrimAge itself reduces death rates.