Epigenetic clocks are among the most significant developments in longevity science in the past decade — not because they solve aging, but because they provide the first validated tools to measure biological aging rate in a living person from a blood or saliva sample. Before these clocks, longevity research had no reliable surrogate endpoint: you had to wait for disease or death. Epigenetic clocks provide a quantifiable, actionable measurement of biological aging that responds to interventions within months rather than decades.
DNA methylation is the mechanism. As cells age, specific cytosine residues in CpG dinucleotides throughout the genome gain or lose methyl groups in highly predictable patterns. These patterns are so consistent across individuals that a machine learning algorithm trained on thousands of samples can predict chronological age with near-perfect accuracy from methylation data alone — and deviations from the expected pattern predict health outcomes better than chronological age.
| Clock | Developed | CpG Sites | Trained On | Best Use Case | Limitation |
|---|---|---|---|---|---|
| Horvath Clock | 2013 | 353 | Chronological age (51 tissues) | Pan-tissue biological age baseline; most validated and cited | Less predictive of mortality than later clocks; tissue-specific variations exist |
| Hannum Clock | 2013 | 71 | Chronological age (blood) | Blood-specific biological age; correlates with telomere length | Blood-specific only; less commonly used in commercial tests |
| PhenoAge (Levine) | 2018 | 513 | Clinical biomarkers (albumin, creatinine, glucose, CRP, etc.) | Phenotypic age reflecting metabolic and inflammatory status; bridges blood biomarkers and methylation | More influenced by acute illness and inflammation — can fluctuate with short-term health changes |
| GrimAge (Lu) | 2019 | 1,000+ | Plasma proteins + time-to-death (mortality) | Best mortality prediction; most sensitive to smoking and cardiovascular risk; gold standard for longevity research | Less responsive to short-term interventions; designed for long-term risk stratification |
| DunedinPACE | 2022 | 173 | Longitudinal biomarker trajectories (pace of aging) | Tracking intervention response; measuring whether lifestyle changes are slowing aging rate; "speedometer" not "odometer" | Newer, fewer validation studies outside Dunedin cohort; less familiar to clinicians |
| Intervention | Effect on Biological Age | Study |
|---|---|---|
| Aerobic exercise (sustained) | −3 to −4 years (Horvath clock) | Dunn 2019: master athletes (lifelong endurance training) showed biological age 4 years younger vs sedentary controls after controlling for BMI; exercise is the most consistently replicated biological age reducer across multiple clock types |
| Caloric restriction (20–40% CR) | −2 to −3 years (GrimAge) | CALERIE 2 trial: 2-year 25% CR → significant GrimAge deceleration; DunedinPACE decreased 2–3% (slowing aging pace by 2–3%); effect correlates with degree of restriction; most studied longevity intervention in humans |
| Mediterranean diet | −1 to −2 years | Dolinoy 2007; multiple cohort studies; Hosseini 2022 meta-analysis: Mediterranean diet adherence associated with lower epigenetic age across multiple clock types; mechanism: polyphenols, omega-3, fiber → reduced oxidative stress and inflammation |
| Smoking cessation | +4.3 years (if currently smoking) → normalizes after 5 years cessation | Smoking is the strongest accelerator of epigenetic aging across all clocks; GrimAge was partially trained on smoking history; biological age returns toward non-smoker levels within 5 years of cessation |
| Sleep (7–9 hours, good quality) | −1 to −2 years vs poor sleepers | Carroll 2022: short sleep duration (<6 hours) and poor sleep quality both associated with GrimAge acceleration; mechanism: sleep deprivation → increased inflammatory markers → methylation changes; sleep is the only time-gated longevity intervention — cannot be "made up" |
| Obesity (BMI >30) | +2 to +4 years acceleration | Multiple cohort studies; adipose tissue inflammation → systemic inflammatory methylation changes; weight loss reverses a portion of biological age acceleration proportional to weight lost |
| Fasting-mimicking diet (FMD) | −2.5 years (Levine PhenoAge) | Brandhorst 2024 (Nature Communications, N=100 RCT): 3 monthly FMD cycles → 2.5-year reduction in biological age, reduced risk factors for metabolic disease; effect persisted after returning to normal diet between cycles |
TruAge Complete (TruDiagnostic): Most comprehensive consumer test; measures Horvath, GrimAge, DunedinPACE, and additional clocks from a dried blood spot (finger prick at home); includes detailed report with biological age, pace of aging, organ-specific age estimates, and lifestyle factor analysis; ~$299–$350; run every 6–12 months to track interventions; the market leader in direct-to-consumer epigenetic testing.
Elysium Index: Uses saliva sample (more convenient); measures a proprietary methylation index; less comprehensive than TruDiagnostic; lower price point (~$199); good entry-level option but does not include DunedinPACE.
Tally Health (DunedinPACE): Subscription model; includes DunedinPACE specifically; paired with coaching and lifestyle tracking; good choice if tracking pace of aging response to interventions is the primary goal.
Interpreting your results: A single test has limited value — biological age measurements have meaningful measurement variability (test-retest variation of ±1–2 years); the value is in tracking change over time after sustained interventions; allow minimum 6 months between tests for meaningful signal; variables that acutely inflate methylation age (recent illness, major surgery, extreme stress, alcohol binge) should be avoided 4–6 weeks before testing for a "clean" baseline; the delta (change from your personal baseline) is more meaningful than the absolute number.
As an Amazon Associate, LongevityLab earns from qualifying purchases made through links on this page. This does not affect the price you pay.