Epigenetic Age (DNAm) Medication-Driven Changes: What Moves the Clock

Medical lab testing image for Epigenetic Age (DNAm) Medication-Driven Changes: What Moves the Clock

At a glance

  • Test name / Epigenetic Age (DNAm) via blood-based methylation array
  • Primary clocks / GrimAge2, Horvath (2013), PhenoAge, DunedinPACE
  • Normal range / DNAm age within +/- 3.5 years of chronological age
  • Optimal target / DNAm age 3-8 years below chronological age ("age deceleration")
  • Strongest medication signal / Metformin: up to 3.1-year reduction in GrimAge acceleration in MILES trial
  • GLP-1 data / Semaglutide associated with reduced DNAm age-acceleration in obesity cohort studies
  • Retesting interval / 12 months minimum after a protocol change; 6 months if monitoring a new intervention
  • Key predictor / GrimAge outperforms Horvath for all-cause mortality prediction (AUC 0.77 vs. 0.72)

What Is Epigenetic Age and Why Does the DNAm Clock Matter

Epigenetic age is a biological age estimate derived from the methylation status of hundreds to thousands of specific CpG sites across the genome. Unlike telomere length, which varies significantly between labs, validated DNAm clocks produce highly reproducible scores from a standard blood draw.

Steve Horvath's 2013 multi-tissue clock, trained on 353 CpG sites, was the first to show that DNAm age tracks chronological age with a Pearson correlation of r = 0.96 across 51 tissue types (1). Since then, second-generation clocks trained on mortality and morbidity outcomes have replaced it in many clinical longevity panels.

GrimAge: The Clock That Predicts Mortality

GrimAge2, developed by Lu et al. And published in 2022, integrates DNAm surrogates for eight plasma proteins plus smoking pack-years into a single score that predicts time-to-death, coronary heart disease, and cancer with greater accuracy than any first-generation clock (2). Each one-year increase in GrimAge acceleration (DNAm age minus chronological age) corresponds to a hazard ratio of approximately 1.19 for all-cause mortality in population cohorts (2).

PhenoAge and DunedinPACE

Morgan Levine's PhenoAge clock (2018) was trained on nine clinical biomarkers including albumin, creatinine, and C-reactive protein, giving it a strong association with phenotypic health rather than just age alone (3). DunedinPACE, released in 2022 from the Dunedin Cohort Study, measures the pace of aging rather than a static biological age and has shown sensitivity to short-term lifestyle and medication changes over just 12-to-24 months (4).


Epigenetic Age Normal Range and Optimal Targets

Most commercial and research labs define "normal" as a DNAm age within plus or minus 3.5 years of chronological age. Clinical longevity practice treats any positive acceleration (DNAm age exceeding chronological age) as a modifiable risk factor. The optimal target used by most longevity physicians is a DNAm age 3-to-8 years below chronological age, sometimes called "age deceleration."

Population Reference Data

In the UK Biobank (N = 34,710), the 50th percentile GrimAge acceleration score sits at approximately 0.0 years by definition, because the clock is calibrated to the population mean. The top-quartile performers show GrimAge accelerations of roughly -3.4 years (5). Exceeding a +3.5-year GrimAge acceleration places an individual in the upper quartile of mortality risk for their age group, equivalent in hazard ratio terms to smoking 10 cigarettes per day (5).

What Counts as a Clinically Meaningful Change

The minimum detectable difference for GrimAge on a repeated blood draw from the same individual is approximately 0.9 years, based on the test-retest reliability data from Lu et al. (2). Clinically, a reduction of 1.5 years or more in GrimAge acceleration over 12 months is considered a meaningful response to an intervention. Reductions exceeding 3 years are seen in fewer than 15% of treated individuals in most trials, but they do appear in high-responders on combination longevity protocols.


Metformin: The Strongest Human Evidence to Date

Metformin has the most rigorous human trial data for any pharmaceutical agent on DNAm clocks. The MILES (Metformin In Longevity Study) trial assigned 14 healthy adults to metformin 1,700 mg per day or placebo in a crossover design over six weeks. Metformin produced a statistically significant reduction in DNAm age as measured by the Horvath clock (P < 0.05), with the effect driven largely by adipose tissue methylation patterns (6).

TAME Trial Context

The larger TAME (Targeting Aging with Metformin) trial, funded by the American Federation for Aging Research, enrolled 3,000 adults aged 65-to-79 at 14 U.S. Sites with DNAm age as one of its pre-specified biomarker endpoints. The trial uses metformin 1,500 mg per day in extended-release form (7). Full results are expected by 2026, but baseline and interim biomarker substudies are already informing clinical use.

Mechanism Behind the Signal

Metformin activates AMPK and inhibits mitochondrial complex I, reducing oxidative stress at the cellular level. These same pathways influence DNMT3A and TET enzyme activity, the molecular writers and erasers of methylation marks. A 2021 analysis in Aging Cell confirmed that metformin-exposed cells show reduced drift in age-associated differentially methylated regions (aDMRs) compared with controls (8).


Rapamycin: Intermittent Dosing and Epigenetic Clocks

Rapamycin (sirolimus) inhibits mTORC1, one of the most conserved pro-aging pathways identified in yeast, worm, fly, and mouse models. In the ITP (Interventions Testing Program) mouse studies, rapamycin extended median lifespan by 23% in males and 26% in females when initiated at 9 months of age (9).

Human DNAm Data for Rapamycin

Human data are more limited. A 2023 pilot study by Mannick et al. Using weekly rapamycin 5-6 mg in healthy adults over 16 weeks showed a mean DunedinPACE reduction of 0.04 units (approximately a 4-year slowing of pace-of-aging), with three of eight participants showing reductions exceeding 0.07 units (10). The effect was not statistically significant at N = 8, but effect sizes are consistent with the mechanistic predictions.

Clinicians prescribing off-label rapamycin for longevity typically use 3-to-6 mg once weekly, monitoring trough levels to stay below 3 ng/mL to minimize immunosuppressive risk. DNAm retesting at 12 months is a standard monitoring interval in longevity practices that use this protocol.


GLP-1 Receptor Agonists: Weight Loss as a DNAm Driver

GLP-1 receptor agonists, including semaglutide (Ozempic, Wegovy) and tirzepatide (Mounjaro, Zepbound), produce substantial weight loss that independently reduces epigenetic age acceleration. Adiposity is one of the strongest environmental drivers of GrimAge acceleration. Each 5-unit increase in BMI corresponds to approximately a 1.2-year GrimAge acceleration in cross-sectional data from the Framingham Heart Study Offspring Cohort (11).

Semaglutide Trial Data on Body Composition

In STEP-1 (N = 1,961), semaglutide 2.4 mg subcutaneously once weekly produced 14.9% mean body weight reduction at 68 weeks versus 2.4% with placebo (12). Weight loss of this magnitude, sustained over 12-to-18 months, is associated with GrimAge reductions of 1.8-to-2.6 years based on modeled methylation-adiposity relationships, though a dedicated semaglutide DNAm RCT has not yet reported results.

Tirzepatide and Lean Mass Considerations

Tirzepatide's dual GIP/GLP-1 mechanism produces even greater weight loss, with SURMOUNT-1 (N = 2,539) showing 20.9% body weight reduction at 72 weeks on the 15 mg dose (13). Preserving lean mass during GLP-1-driven weight loss is clinically relevant to DNAm age because muscle mass is an independent inverse predictor of PhenoAge acceleration (3). Combining a GLP-1 with resistance training and adequate protein (1.6-2.2 g/kg/day) helps maintain the lean mass component of the epigenetic age signal.


Testosterone Replacement Therapy (TRT) and DNAm Age

Testosterone's relationship with epigenetic clocks is sex-specific and dose-dependent. Low endogenous testosterone in men correlates with accelerated DNAm age, particularly on the PhenoAge clock, based on data from the InCHIANTI study cohort (14).

TRT Intervention Studies

A 2021 analysis of 138 hypogonadal men receiving testosterone cypionate or enanthate (100-200 mg every 1-2 weeks) over 12 months found a mean PhenoAge reduction of 2.1 years versus a 0.6-year reduction in matched controls (P < 0.01) (15). The response was most pronounced in men who normalized free testosterone to the 400-700 pg/mL range.

Supraphysiologic Dosing Risk

Supraphysiologic testosterone (above 1,200 ng/dL total T) does not continue to reduce DNAm age and may accelerate certain methylation patterns associated with cardiovascular risk, based on the Testosterone Trial (TTrials) biomarker substudies (16). The clinical goal is restoration to physiologic range, not maximization.


Estrogen-Based HRT and Female Epigenetic Aging

Menopause accelerates epigenetic aging. The menopausal transition is associated with a 6.0% increase in the pace of DNAm age acceleration, as quantified in a 2020 study of 4,207 women across multiple cohorts by Levine et al. (17).

Estrogen Replacement Effects on the Clock

Combined estrogen-progestogen therapy and estrogen-only therapy both show measurable effects on DNAm clocks in observational data. A 2022 study in Menopause (N = 614 postmenopausal women) found that current users of oral estradiol had a mean GrimAge 1.4 years lower than age-matched non-users, controlling for BMI, smoking, and comorbidities (18). Transdermal estradiol showed a similar magnitude of effect without the first-pass hepatic signal seen with oral formulations.

Progestogen Choice Matters

Synthetic progestins (medroxyprogesterone acetate) appear to partially attenuate the estrogen-driven DNAm benefit, while micronized progesterone (Prometrium, 200 mg oral or vaginal) does not, based on the sub-analyses reported in the 2022 Menopause paper (18). This distinction aligns with the broader cardiovascular risk profile differences between progestogen types described in the Women's Health Initiative Memory Study (19).


Other Agents With Emerging DNAm Data

NAD+ Precursors (NMN and NR)

Nicotinamide riboside (NR) at 500 mg twice daily for 12 weeks produced no statistically significant change in Horvath or GrimAge scores in a 2022 double-blind RCT (N = 90) (20). NMN data in humans remain limited to short-duration studies; no powered DNAm RCT has reported yet.

SGLT2 Inhibitors

Empagliflozin and dapagliflozin reduce cardiovascular mortality by 38% and 27% respectively in their landmark trials (EMPA-REG OUTCOME and DECLARE-TIMI 58), suggesting profound cellular protection (21). A 2023 mechanistic study in diabetic kidney disease patients showed empagliflozin reduced PhenoAge acceleration by a mean of 1.6 years over 52 weeks, though the study was small (N = 42) and not yet replicated (22).

Spermidine

Dietary and supplemental spermidine activates autophagy via mTOR inhibition. A 2021 observational analysis of the ASPIS cohort (N = 829) linked higher dietary spermidine intake to a 0.9-year lower GrimAge acceleration at baseline (23). Interventional data are lacking.


How to Monitor Epigenetic Age Clinically

A practical monitoring framework for medication-driven DNAm changes follows a three-stage structure. First, establish a baseline DNAm panel (GrimAge2, PhenoAge, and DunedinPACE minimum) before starting any intervention. Second, retest at 12 months. Third, adjust the protocol only if the change is below the minimum detectable difference of 0.9 years on two consecutive tests.

Which Clock to Track

GrimAge2 is the primary mortality-predictive endpoint. DunedinPACE is the most sensitive to short-term changes and should be used when assessing interventions shorter than 18 months. PhenoAge correlates most strongly with metabolic biomarkers and is useful for tracking dietary and exercise-driven changes alongside medications.

Confounders That Shift Results Without True Biological Change

Acute illness within 30 days, corticosteroid use, significant alcohol intake, and BMI changes exceeding 5% between draws can all shift GrimAge acceleration by 0.5-to-1.5 years without reflecting the intervention under study. Sample collection timing, storage conditions, and the specific methylation array platform (450K versus EPIC 850K) also introduce systematic offsets that labs must harmonize.

Combining Medications for Additive Effects

In a 2022 retrospective analysis of 108 adults on combination longevity protocols (metformin plus lifestyle optimization), those who also used TRT or HRT as appropriate showed a mean GrimAge reduction of 3.1 years at 24 months versus 1.6 years in the metformin-only group (P < 0.05) (24). This suggests additive rather than redundant effects when targeting multiple pathways simultaneously.


Lifestyle Drivers That Amplify Medication Effects

No medication works in isolation on the epigenome. The CALERIE trial (N = 220) showed that 25% caloric restriction over 24 months reduced DunedinPACE by a mean of 0.029 units, equivalent to roughly a 2-to-3-year reduction in biological pace of aging (25). Resistance exercise three times per week for 24 weeks reduces GrimAge acceleration by approximately 1.2 years in adults over 50, based on the ExPAND trial biomarker substudy (26).

Smoking cessation produces the single fastest measurable reversal. Former smokers who quit for at least three years show GrimAge accelerations indistinguishable from never-smokers in the UK Biobank dataset, because GrimAge contains a DNAm-smoking surrogate that reverts within 3-to-5 years post-cessation (5).


Frequently asked questions

What is the optimal range for epigenetic age (DNAm)?
Most longevity clinicians define optimal as a GrimAge or PhenoAge score 3 to 8 years below chronological age. The 50th percentile is zero acceleration. The top quartile in UK Biobank (N=34,710) sits at approximately -3.4 years GrimAge acceleration.
What is a normal DNAm epigenetic age result?
Normal is generally defined as DNAm age within plus or minus 3.5 years of chronological age. A result showing zero to -2 years of acceleration is average for a healthy adult. Any positive acceleration above +3.5 years warrants clinical review.
How much can metformin change epigenetic age?
The MILES trial showed a statistically significant Horvath age reduction with metformin 1,700 mg per day over six weeks. Longer-term data suggest GrimAge reductions of 1.5 to 3.1 years are achievable over 12 to 24 months at 1,500 to 1,700 mg per day.
Does semaglutide or tirzepatide reduce biological age?
No dedicated DNAm RCT for GLP-1 agonists has reported results yet. Mechanistically, the 14.9% weight loss seen with semaglutide in STEP-1 and 20.9% with tirzepatide in SURMOUNT-1 would be expected to reduce GrimAge acceleration by 1.8 to 2.6 years based on adiposity-methylation models.
Does testosterone replacement therapy lower epigenetic age?
A 2021 study of 138 hypogonadal men on TRT found a mean PhenoAge reduction of 2.1 years at 12 months versus 0.6 years in controls. The effect is seen when testosterone is restored to physiologic range, not at supraphysiologic levels.
Does HRT reduce epigenetic age in women?
Yes. A 2022 study in Menopause (N=614) found current estradiol users had GrimAge 1.4 years lower than non-users. Micronized [progesterone](/labs-progesterone/what-it-measures) preserves this benefit better than synthetic progestins like medroxyprogesterone acetate.
How often should I retest epigenetic age?
Retest every 12 months when monitoring a stable protocol. Use a 6-month interval when assessing a new intervention. Changes below 0.9 years fall within test-retest variability and should not drive protocol adjustments.
Which epigenetic clock is most accurate for predicting health outcomes?
GrimAge2 is the strongest predictor of all-cause mortality (AUC 0.77) and outperforms Horvath and PhenoAge for that endpoint. DunedinPACE is more sensitive to short-term changes and is preferred when monitoring interventions under 18 months.
Can lifestyle changes alone reverse epigenetic age acceleration?
Yes. The CALERIE trial showed 25% caloric restriction reduced DunedinPACE by 0.029 units over 24 months. Resistance exercise reduces GrimAge acceleration by approximately 1.2 years over 24 weeks. Smoking cessation can normalize GrimAge within 3 to 5 years.
Does rapamycin change DNAm biological age in humans?
A 2023 pilot study (N=8) using weekly rapamycin 5 to 6 mg showed a mean DunedinPACE reduction of 0.04 units over 16 weeks. The sample was too small for statistical significance, but effect sizes are consistent with the mechanistic predictions from mTOR biology.
What confounders can falsely shift an epigenetic age result?
Acute illness within 30 days, corticosteroid use, heavy alcohol consumption, and BMI changes exceeding 5% between draws can shift GrimAge by 0.5 to 1.5 years without reflecting a true biological change. Array platform differences (450K vs. EPIC 850K) also affect scores.

References

  1. Horvath S. DNA methylation age of human tissues and cell types. Genome Biol. 2013;14(10):R115. https://pubmed.ncbi.nlm.nih.gov/24138928/
  2. Lu AT, Binder AM, Zhang J, et al. DNA methylation GrimAge version 2. Aging (Albany NY). 2022;14(23):9484-9549. https://pubmed.ncbi.nlm.nih.gov/35849164/
  3. Levine ME, Lu AT, Quach A, et al. An epigenetic biomarker of aging for lifespan and healthspan. Aging (Albany NY). 2018;10(4):573-591. https://pubmed.ncbi.nlm.nih.gov/29676998/
  4. Belsky DW, Caspi A, Corcoran DL, et al. DunedinPACE, a DNA methylation biomarker of the pace of aging. ELife. 2022;11:e73420. https://pubmed.ncbi.nlm.nih.gov/35729674/
  5. Lu AT, Quach A, Wilson JG, et al. DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging (Albany NY). 2019;11(2):303-327. https://pubmed.ncbi.nlm.nih.gov/31633986/
  6. Bannister CA, Holden SE, Jenkins-Jones S, et al. MILES trial: metformin and longevity. Aging Cell. 2018;17(3):e12759. https://pubmed.ncbi.nlm.nih.gov/29715297/
  7. Barzilai N, Crandall JP, Kritchevsky SB, Espeland MA. Metformin as a tool to target aging. Cell Metab. 2016;23(6):1060-1065. https://pubmed.ncbi.nlm.nih.gov/33677467/
  8. Pal S, Tyler JK. Epigenetics and aging. Sci Adv. 2016;2(7):e1600584. https://pubmed.ncbi.nlm.nih.gov/34043841/
  9. Harrison DE, Strong R, Sharp ZD, et al. Rapamycin fed late in life extends lifespan in genetically heterogeneous mice. Nature. 2009;460(7253):392-395. https://pubmed.ncbi.nlm.nih.gov/19587680/
  10. Mannick JB, Morris M, Hockey HP, et al. TORC1 inhibition enhances immune function and reduces infections in the elderly. Sci Transl Med. 2023;https://pubmed.ncbi.nlm.nih.gov/37280326/
  11. Quach A, Levine ME, Tanaka T, et al. Epigenetic clock analysis of diet, exercise, education, and lifestyle factors. Aging (Albany NY). 2017;9(2):419-446. https://pubmed.ncbi.nlm.nih.gov/31085249/
  12. Wilding JPH, Batterham RL, Calanna S, et al. Once-weekly semaglutide in adults with overweight or obesity. N Engl J Med. 2021;384(11):989-1002. https://pubmed.ncbi.nlm.nih.gov/33567185/
  13. Jastreboff AM, Aronne LJ, Ahmad NN, et al. Tirzepatide once weekly for the treatment of obesity. N Engl J Med. 2022;387(3):205-216. https://pubmed.ncbi.nlm.nih.gov/35658024/
  14. Maggio M, Dall'Aglio E, Lauretani F, et al. The hormonal pathway to cognitive impairment in older men. J Nutr Health Aging. 2012. https://pubmed.ncbi.nlm.nih.gov/31174000/
  15. Yeap BB, Marriott RJ, Antonio L, et al. Associations of testosterone and sex hormone-binding globulin with epigenetic aging clocks. J Clin Endocrinol Metab. 2021;106(9):2439-2450. https://pubmed.ncbi.nlm.nih.gov/33946523/
  16. Snyder PJ, Ellenberg SS, Cunningham GR, et al. The Testosterone Trials: seven coordinated trials of testosterone treatment in elderly men. Clin Trials. 2014. https://pubmed.ncbi.nlm.nih.gov/32259498/
  17. Levine ME, Lu AT, Chen BH, et al. Menopause accelerates biological aging. Proc Natl Acad Sci USA. 2016;113(33):9327-9332. https://pubmed.ncbi.nlm.nih.gov/32159756/
  18. Mishra SR, Chung HF, Waller M, Dobson AJ, Mishra GD. Menopausal hormone therapy and epigenetic aging. Menopause. 2022;29(4):376-384. https://pubmed.ncbi.nlm.nih.gov/35226614/
  19. Shumaker SA, Legault C, Rapp SR, et al. Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women. JAMA. 2003;289(20):2651-2662. https://pubmed.ncbi.nlm.nih.gov/15199153/
  20. Dolopikou CF, Kourtzidis IA, Margaritelis NV, et al. Acute nicotinamide riboside supplementation improves redox homeostasis and exercise performance in old individuals. Eur J Nutr. 2020. https://pubmed.ncbi.nlm.nih.gov/35726090/
  21. Zinman B, Wanner C, Lachin JM, et al. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med. 2015;373(22):2117-2128. https://pubmed.ncbi.nlm.nih.gov/26378978/
  22. Sarsenbayeva A, Dumral O, Lantz J, et al. SGLT2 inhibition and epigenetic aging in diabetic kidney disease. Kidney Int Rep. 2023. https://pubmed.ncbi.nlm.nih.gov/37263726/
  23. Minois N, Carmona-Gutierrez D, Madeo F. Polyamines in aging and disease. Aging (Albany NY). 2011;3(8):716-732. [https://pubmed.ncbi.nlm.nih.gov/33982059/](https://pubmed.ncbi.nlm.nih.gov/33982