Interventions That Lower Epigenetic Age: What 41 Human Studies Actually Show

At a Glance

The epigenetic clock field has outpaced its own therapeutics. For years, longevity clinics have promised patients that a stack of supplements and intermittent infusions would “reverse biological age.” Then a 2026 systematic review catalogued every published human trial testing next-generation epigenetic clocks and found a more complicated picture: the highest-impact interventions are largely free, the priciest have no measurable effect on the clocks, and one popular premium procedure appears to make things worse.

This is a clinical update that matters. The conclusions reshape how I discuss longevity protocols with patients — and they should inform anyone investing significant time, money, or personal risk in an anti-aging programme.

Why Next-Generation Clocks Change the Calculus

The first-generation Horvath methylation clock (2013) was a proof of concept: DNA methylation patterns in blood could estimate age with remarkable accuracy. Its successors — GrimAge, PhenoAge, DunedinPACE — are built differently. Rather than predicting chronological age, they are trained on mortality and disease outcomes. GrimAge, for instance, predicts time-to-death more accurately than chronological age in multiple large cohorts. DunedinPACE estimates the rate of aging in real time, as a continuous pace-of-change metric.

This distinction matters for clinical decision-making. An intervention that lowers the Horvath clock but does not affect GrimAge or DunedinPACE may be resetting a statistical pattern without slowing the biological processes that actually kill people. The 2026 Johnson & Sinclair review in Frontiers in Genetics (PMID 42294499) deliberately restricted its analysis to these newer, outcome-validated clocks. Of 41 human studies included, the findings diverge sharply from the marketing around many premium longevity interventions.

What Consistently Lowers the Clock

Exercise

The most replicated finding across the entire dataset is physical activity. A 2025 analysis of the Health and Retirement Study (n = large nationally representative cohort, PMID 40511567) found that physically active adults showed significantly lower epigenetic age acceleration on GrimAge, PhenoAge, and DunedinPACE simultaneously. Crucially, both recent activity and cumulative lifetime activity predicted the effect — it is not simply about what you did last month.

The mechanism is multifactorial: exercise reduces circulating inflammatory cytokines (IL-6, TNF-α, CRP), induces mitochondrial biogenesis, stimulates AMPK signalling, and improves insulin sensitivity — all of which feed into the biological processes the clocks are trained on. For patients asking what gives the most clock benefit per unit of time, the answer is consistent: a combination of moderate aerobic exercise (Zone 2, 150+ minutes/week) and resistance training (2–3 sessions/week) produces the strongest combined signal.

Practically, I recommend patients track this with VO2 max as an objective proxy. We now have good data that VO2 max correlates strongly with DunedinPACE independently of weight or other lifestyle factors.

Diet: Methylation-Supportive and Plant-Rich Patterns

The Fitzgerald group has produced the most detailed dietary intervention data. Their 2021 pilot RCT in Aging (PMID 33844651) tested an 8-week programme combining a methylation-supportive diet (folate, B12, betaine, polyphenols), structured sleep, exercise, relaxation guidance, and specific supplements in men aged 50–72. Epigenetic age dropped approximately 3.2 years relative to controls — a striking effect for 8 weeks, though the trial was small (n = 43) and needs replication.

Their 2023 women’s case series (PMID 36947707) found a mean biological age reduction of approximately 4.6 years across six participants using the same protocol.

What is shared across these diets: high intake of dark leafy greens and cruciferous vegetables (sulforaphane and folate), minimisation of ultra-processed foods, adequate protein for muscle maintenance, and polyphenol-dense foods (berries, green tea, olive oil). This is not a novel food pattern — it tracks closely with what we have known from cardiovascular and cancer epidemiology for decades. The epigenetic clock data gives it a new mechanistic frame.

Caloric Restriction

The CALERIE-2 trial — the largest long-term caloric restriction RCT in humans — showed that sustained 25% caloric restriction over two years produced measurable reductions in PhenoAge and other outcome-validated clocks. The Johnson & Sinclair 2022 review in Aging Cell (PMID 35778957) confirmed this pattern across multiple lines of evidence: caloric restriction appears to work partly by reducing inflammation, lowering insulin-like growth factor 1, and activating SIRT1 and AMPK pathways.

For clinical practice, sustained 25% restriction is rarely feasible long-term. Time-restricted eating (16:8) and 5:2 protocols produce smaller but measurable effects in the available data. The key driver appears to be total energy reduction, not fasting duration per se.

GLP-1 Receptor Agonists (Semaglutide)

This is the most surprising finding in the 2026 review. Semaglutide showed a signal toward lower epigenetic age across 1–2 trials with clock endpoints, independent of its weight-loss effect. The proposed mechanism involves GLP-1 receptor expression in the brain, anti-inflammatory properties, and metabolic normalisation beyond simple caloric restriction.

This should be interpreted cautiously — the data are early and the trials were not primarily designed to test epigenetic outcomes. However, given the metabolic and cognitive neuroprotection data (see our article on GLP-1 and Alzheimer’s risk), the signal is biologically plausible.

What Does Not Move the Clock

This is where the 2026 systematic review becomes genuinely important for clinical decision-making, and where I have had to update some of my own prior assumptions.

Nicotinamide Riboside (NR)

NR is among the most widely sold longevity supplements. The mechanistic logic — NAD+ precursor → SIRT1 and PARP1 activity → DNA repair and mitochondrial function — is sound and reproducible in animal models. Human RCTs, however, have consistently failed to show a detectable effect on next-generation epigenetic clocks. This does not mean NR is without biological effect; it may be doing something the clocks are not designed to detect. But patients should not expect their GrimAge or DunedinPACE to improve with NR supplementation based on current evidence.

This is consistent with our more cautious stance on pure NAD precursor supplementation. For patients interested in NAD optimisation, IV NAD+ appears to have acute neurological benefits distinct from what oral NR produces — but neither currently has robust epigenetic clock data showing clock reversal.

Rapamycin

Rapamycin’s mTOR-inhibiting mechanism is among the most compelling in geroscience, and it extends lifespan in every organism model tested. Yet published human trials (mostly low-dose pulsed protocols) have produced no consistent signal in next-generation epigenetic clocks. This may reflect the small sample sizes and short durations of available human data, or it may reflect a genuine disconnect between mTOR-mediated lifespan extension and the biological processes captured by methylation clocks. We do not yet know. The absence of a clock signal does not close the case on rapamycin — it means we need better-powered trials.

Senolytics (Dasatinib + Quercetin)

The dasatinib + quercetin combination — the most studied senolytic protocol — clears senescent cells in human tissue and has produced measurable benefits in idiopathic pulmonary fibrosis and diabetic kidney disease. Clock data, however, are absent or null in published trials. Again, the mechanistic logic remains plausible; the clock evidence is simply not there yet.

Plasmapheresis: A Warning

One finding in the Johnson & Sinclair review deserves special emphasis: plasmapheresis (blood plasma exchange) for longevity purposes showed accelerated epigenetic aging in one trial. This is a single data point and needs replication, but it runs directly counter to the marketing around “young blood” and plasma exchange as anti-aging procedures. Until more data are available, recommending plasma exchange as an anti-aging intervention would not be supported by the best current clock evidence — and given the potential for harm, the bar for evidence should be high.

This does not apply to therapeutic apheresis for specific indications (Lyme disease, autoimmune conditions, hypercholesterolaemia). That is a different mechanism and a different risk-benefit profile.

A Practical Framework: Stacking What Works

Based on the current evidence hierarchy, a rational protocol for patients targeting epigenetic age reduction looks like this:

Tier 1 — Highest evidence, low cost, low risk:

• Consistent aerobic + resistance exercise (Zone 2 + strength training)
• Plant-rich, methylation-supportive diet (dark greens, cruciferous, berries, olive oil)
• Adequate sleep (the Fitzgerald protocol included structured sleep hygiene)
• Caloric moderation (not necessarily severe restriction)

Tier 2 — Meaningful signal, some caveats:

• GLP-1 agonist if metabolic indication exists (do not use for clock benefit alone)
• Metformin if glucose dysregulation is present
• Vitamin D3 optimisation to physiological levels

Tier 3 — Mechanistically plausible, clock evidence absent:

• Rapamycin pulsed protocols (monitor carefully; lack of clock signal does not mean lack of benefit)
• Senolytics (periodic cycles may still be warranted for senescent cell burden)
• NR/NMN supplementation (adjunct role, not primary clock intervention)

The core message for patients: the best-evidenced interventions are not the ones being sold at premium prices. Exercise and diet move the clock more reliably than any supplement or infusion tested so far in humans. Expensive clinic-based protocols may still have value for specific pathways, but they should not be substituted for the basics.

Tracking Progress: Which Test to Use

If you are running an intervention protocol and want to track epigenetic age objectively, TruAge, GlycanAge, and OMICmAge are the current clinical leaders. For monitoring pace of change specifically, DunedinPACE (available via TruDiagnostic’s TruAge Complete) is the most sensitive marker for detecting short-term intervention effects. Run a baseline, hold the protocol for 8–12 weeks minimum before retesting, and change only one or two variables at a time so the signal can be attributed.

Related Articles

• Epigenetic Clock Tests Compared: TruAge, GlycanAge, and OMICmAge
• Rapamycin for Longevity: Dosage, Pulsing, and What We Know
• Senolytics: Clearing Senescent Cells for Longevity
• NAD+ IV Therapy: What to Expect
• Zone 2 Training and VO2 Max for Longevity