What most effectively reduces epigenetic age in humans? An analysis of the most effective interventions

Aug 4, 2026

An analysis of studies examining diets, exercise, supplements, medications, and therapies that influence next-generation epigenetic aging clocks.

What most effectively reduces epigenetic age in humans? An analysis of the most effective interventions

Introduction

Epigenetic clocks are increasingly used to estimate the biological pace of aging. By analyzing patterns of DNA methylation, they attempt to determine whether a person’s biology resembles that of someone younger, older, or aging faster than their chronological age would suggest.

The growing popularity of these tests has created an appealing simplification: if a diet, supplement, or medication lowers the clock result by several years, the body must have been rejuvenated by exactly the same amount.

That conclusion cannot currently be supported.

An epigenetic clock result is a biomarker. It may reflect improved metabolic health, reduced inflammation, better body composition, or changes in immune function. However, it may also be influenced by temporary shifts in blood-cell composition, stress, infection, strenuous exercise, the timing of sample collection, and the characteristics of the algorithm itself.

A systematic review of human intervention studies collected experiments that used next-generation epigenetic clocks to assess the effects of diet, calorie restriction, physical activity, supplements, metabolic medications, psychiatric treatments, clinical procedures, and experimental longevity interventions.

The main objective was not merely to identify methods that changed a laboratory result. The more important question was which interventions produced the most consistent and biologically plausible signals of slower epigenetic aging.

What is epigenetic age?

DNA methylation is one of the mechanisms involved in regulating gene activity. It includes the attachment of methyl groups to specific locations in DNA. This does not alter the genetic sequence itself, but it can influence which genes are more active and which remain relatively suppressed.

As people age, some methylation patterns change in relatively predictable ways. This made it possible to create algorithms that estimate biological age using samples of blood, saliva, or other tissues.

The earliest epigenetic clocks were designed primarily to predict chronological age as accurately as possible. If an algorithm analyzing the DNA of a 50-year-old produced a result close to 50, it was considered accurate.

Newer clocks have a broader purpose. Their results are intended to reflect not only the number of years lived, but also factors associated with:

Chronic disease risk, because some clocks incorporate methylation patterns linked to metabolic and cardiovascular health.
Metabolic function, including signals associated with glucose regulation, lipid metabolism, and inflammatory activity.
Physical capacity, because certain models attempt to capture aspects of functional aging rather than age alone.
Disability risk, particularly when the clock was trained using clinical outcomes associated with loss of independence.
Premature mortality, which is a central target of models such as GrimAge.

Frequently used models include DNAm PhenoAge, GrimAge, GrimAge2, DunedinPACE, OMICmAge, and FitAge.

These clocks do not all measure the same thing. DNAm PhenoAge is more strongly connected with clinical features associated with aging. GrimAge was developed to improve the prediction of disease and mortality. DunedinPACE does not express biological age in years but instead estimates the current pace of biological change.

A DunedinPACE result of 1 broadly represents an average pace of aging. A result of 1.05 suggests that biological changes are occurring at a pace corresponding to approximately 1.05 biological years per chronological year. A result below 1 suggests a slower pace.

Two people may therefore have a similar estimated epigenetic age but a different current pace of aging. One person may appear biologically older while currently aging more slowly, while another may have a more favorable starting point but a worsening rate of change.

Epigenetic age and the epigenetic pace of aging are therefore related but distinct measurements.

Study details

The analysis was conducted as a systematic review of human intervention studies. Its purpose was to identify experiments testing whether specific interventions can alter next-generation epigenetic aging clocks, which are more closely associated with health outcomes, age-related disease, and mortality risk than earlier clocks designed mainly to predict chronological age.

The final review included 41 unique studies in which epigenetic age or the pace of epigenetic aging was measured both before and after the intervention.

The authors analyzed clocks including:

DNAm PhenoAge, which combines DNA methylation patterns with clinical indicators of health and mortality risk.
GrimAge and GrimAge2, designed to predict disease risk and premature mortality.
DunedinPACE and DunedinPoAm, which estimate the pace of aging rather than biological age expressed in years.
FitAge, which incorporates signals associated with physical fitness.
OMICmAge, which uses epigenetic proxies for metabolic, protein-related, and clinical parameters.
CausAge, AdaptAge, DamAge, InflammAge, and Systems Age, which capture partly different biological dimensions of aging.

Studies were included only if they:

Involved an intervention conducted in humans.
Used at least one next-generation epigenetic aging clock.
Included measurements before and after the intervention.
Were published from 2018 onward, when the first clocks of this generation became available.

The review excluded observational and cross-sectional studies, animal experiments, studies conducted only in cells, literature reviews, and papers using only first-generation clocks.

The interventions were divided into four main groups:

Pharmacological interventions, including semaglutide, statins, metformin, ketamine, and antiretroviral therapy.
Lifestyle and supplementation interventions, including physical activity, calorie restriction, plant-rich diets, omega-3 fatty acids, vitamin D, and multivitamin-mineral supplements.
Non-pharmacological clinical and psychosocial interventions, such as kidney transplantation, umbilical cord plasma products, and programs supporting parent-child relationships.
Interventions without a significant effect, including nicotinamide riboside, rapamycin, senolytics, folic acid, and some dietary and exercise programs.

A result was considered significant only when it was present at the end of the full intervention and met the predefined statistical threshold. If an improvement appeared during the study but was no longer present at the final measurement, it was not classified as a successful outcome.

The authors also evaluated sample size, intervention duration, the clocks used, and the magnitude of change. However, the studies varied substantially in quality and scale. Some included more than 100 participants and lasted several years, while others were conducted for only a few weeks in groups of fewer than a dozen people.

The review was therefore not a meta-analysis capable of producing a reliable ranking of interventions. It was primarily a structured overview of interventions that had decreased, failed to change, or increased the result of at least one next-generation epigenetic aging clock.

Source: A comprehensive list of interventions that decrease next-generation epigenetic aging clocks in humans

How were the interventions evaluated?

Only human studies in which an epigenetic clock was measured both before and after an intervention were included.

This made it possible to assess whether a lifestyle change, supplement, medication, or therapy altered the result within the same participant.

Observational studies that merely compared people with different lifestyles were not included. Such studies can reveal associations, but they cannot easily establish whether a particular intervention caused the difference.

The review also excluded:

Animal experiments, because findings in mice or other organisms cannot be directly translated into human effects.
Cell-only studies, because changes in isolated cells do not necessarily reflect whole-body aging.
Studies without before-and-after measurements, because they do not show how an intervention changed the same individual.
Studies using only older clocks designed mainly to predict chronological age.

This matters because some popular claims about reversing biological age come from small uncontrolled studies or from clocks that were not designed to predict disease, functional decline, or mortality.

The interventions were grouped into several broad categories:

Diet and energy restriction.
Physical activity.
Supplementation.
Medications.
Clinical interventions.
Psychosocial programs.
Interventions without statistically significant effects.

Being classified as a positive intervention did not mean that a method had been proven to be an anti-aging therapy.

The strength of the evidence also depended on:

The number of participants, because small studies produce less stable estimates.
The duration of the intervention, because epigenetic changes may require months or years to become detectable.
The presence of a control group, which helps distinguish a real effect from natural variation.
The type of clock used, since different models capture different aspects of aging.
The number of models analyzed, because testing many clocks increases the chance of a random positive finding.
The magnitude of the change, not merely whether it crossed a statistical threshold.
The health status of the participants, since people with greater metabolic dysfunction may have more room for improvement.
Whether the result was reproduced independently.

An intervention that improved one out of ten clocks in a group of fewer than 20 people provides much weaker evidence than a method tested in a large randomized trial that also improved established health markers.

Which interventions worked best?

There is not yet a single intervention that can confidently be described as the most effective way to reduce epigenetic age.

The most promising results appeared in several categories:

Long-term reduction of excess calorie intake.
Diets rich in minimally processed plant foods.
Regular physical activity.
Omega-3 supplementation.
Combined omega-3, vitamin D, and exercise interventions.
Treatments that improved metabolic control and reduced excess body fat.
Selected medications used in people with specific medical conditions.

The most practically convincing interventions are those that not only improved one or more epigenetic clocks, but also have independent evidence supporting their effects on metabolic, cardiovascular, or functional health.

Using that standard, the strongest candidates include:

Regular exercise, because it improves cardiorespiratory fitness, insulin sensitivity, vascular function, muscle mass, and inflammatory regulation.
Avoiding chronic excess energy intake, because it lowers the risk of visceral obesity, insulin resistance, and prolonged metabolic overload.
A nutrient-dense diet, because it influences several biological pathways associated with aging at the same time.
Omega-3 fatty acids, especially when dietary intake is low, because several studies found small but recurring benefits.
Treatment of genuine metabolic disorders, because improvements in glucose control, lipid levels, body weight, and inflammation may be reflected by epigenetic clocks.

The response is unlikely to be identical in every person. The more unfavorable the starting point, the greater the potential for improvement.

Someone with obesity, insulin resistance, low activity, and insufficient omega-3 intake may show a larger response than a lean, physically active person with a high-quality diet.

Calorie restriction and diet

One of the strongest studies evaluated long-term calorie restriction in people without obesity. Participants reduced their energy intake for approximately two years, after which several epigenetic clocks were reassessed.

The clearest result involved DunedinPACE. Calorie restriction was associated with a slower estimated pace of aging.

This did not mean that participants became biologically younger by a specific number of years. The result primarily concerned the rate of ongoing biological change, rather than the reversal of all previous aging processes.

Potential mechanisms include:

Improved insulin sensitivity, which reduces the need for chronically elevated insulin and may lower metabolic strain.
Reduced visceral fat, which can decrease inflammatory signaling and improve organ function.
Lower chronic inflammation, which may influence several methylation patterns included in newer clocks.
Improved lipid profiles, reducing cardiovascular burden.
Reduced activation of growth-related pathways, although the practical significance of this mechanism in humans remains uncertain.

Calorie restriction also has important limitations. An excessive deficit may lead to muscle loss, poorer recovery, reduced libido, hormonal disruption, persistent fatigue, and insufficient nutrient intake.

The goal should not be to consume as few calories as possible. A more sensible target is to avoid long-term energy excess while preserving muscle, strength, performance, and adequate nutrient availability.

Positive results were also observed with diets rich in minimally processed plant foods. Such dietary patterns can increase the intake of:

Fiber, which supports glucose control, satiety, bowel function, and microbial fermentation.
Polyphenols, which may influence inflammatory and cellular stress pathways.
Vitamins and minerals, which support normal metabolic and enzymatic processes.
Potassium, which is relevant to blood-pressure regulation.
Unsaturated fats, which can improve the overall lipid profile when they replace less favorable fat sources.
Compounds that support gut-microbiome diversity.

At the same time, these diets often reduce the consumption of ultra-processed foods, excess sodium, refined sugars, and less favorable sources of dietary fat.

These findings should not be interpreted as proof that every animal-derived food accelerates aging. The total dietary pattern, food quality, energy balance, and individual requirements matter more than a single food category.

A practical eating pattern may include:

A large variety of vegetables.
Regular fruit intake.
Legumes.
Whole grains.
Nuts and seeds.
Fish.
Sufficient protein.
Limited ultra-processed food.

Such a pattern can improve not only epigenetic measurements but also health parameters with much stronger clinical validation.

Physical activity

Regular exercise was associated with favorable signals in some studies, although the results differed between experiments.

In several trials, endurance training reduced the result of at least one epigenetic aging clock. Beneficial changes were also observed in programs combining exercise, nutritional support, and assistance for people with frailty.

Other studies found no statistically significant effect.

This does not mean that exercise fails to slow important aspects of aging. A more likely explanation is that epigenetic clocks do not capture every beneficial adaptation or may require a longer period before detecting meaningful change.

Physical activity affects multiple biological systems simultaneously:

Cardiorespiratory fitness improves the body’s ability to deliver and use oxygen, increasing exercise tolerance and physiological reserve.
Insulin sensitivity improves because active muscles remove glucose from the bloodstream more effectively.
Muscle mass and strength support independence, glucose regulation, and protection against frailty later in life.
Inflammation may decrease through reductions in visceral fat and regular activation of anti-inflammatory signals associated with exercise.
Vascular function improves through increased blood flow and repeated stimulation of the endothelium.
Nervous-system regulation can improve, supporting better sleep, mood, stress tolerance, and cognitive function.

Some of these adaptations may affect DNA methylation, but not every improvement must lead to a lower result on a specific epigenetic clock.

For this reason, exercise effectiveness should not be judged only through biological-age testing. A more useful picture comes from combining:

VO₂max or a standardized endurance test.
Grip strength.
Resistance-training performance.
Muscle mass.
Resting heart rate.
Blood pressure.
Recovery quality.
Daily functional capacity.

A balanced longevity-oriented program should include endurance exercise, resistance training, regular low-intensity movement, mobility, and balance work.

Omega-3, vitamin D, and supplements

Omega-3 fatty acids were among the most consistent supplementation interventions associated with changes in epigenetic aging clocks.

Across several studies, supplementation was linked to favorable changes in at least one model. The reported effects involved clocks such as GrimAge, DNAm PhenoAge, PC PhenoAge, DamAge, and DunedinPACE.

One large study in older adults evaluated:

Omega-3 fatty acids.
Vitamin D.
A home-based exercise program.

These interventions were tested both separately and in combination.

Omega-3 produced the most consistent signal. Combining all three interventions provided an additional benefit in some analyses.

The magnitude of the changes was moderate. There was no dramatic reversal of biological age. The findings instead suggested that small effects may accumulate over several years.

Omega-3 fatty acids may act through:

Changes in cell-membrane composition.
Modulation of inflammatory mediators.
Improvements in lipid metabolism.
Regulation of immune function.
Support for vascular health.
Effects on liver metabolism.

The benefit may be greater in people who consume little oily fish or have low levels of EPA and DHA.

This does not mean that high doses should be used by everyone. Supplement quality, EPA and DHA content, oxidation, medications, bleeding risk, and cardiovascular status should all be considered.

Some studies also reported favorable changes after multivitamin-mineral supplementation. The effect may have resulted primarily from correcting small nutritional deficiencies in older adults.

When a nutrient deficiency limits normal biological function, correcting it may improve several systems. Someone without a deficiency may experience a much smaller benefit or none at all.

This distinction is important: supplementation is most likely to help when it addresses a genuine limitation, not when it is added to an already well-optimized lifestyle without a clear reason.

Several popular longevity supplements did not produce convincing effects, including some NAD+ precursors, polyphenol blends, and multi-ingredient formulations.

Metabolic medications

Some of the largest changes in epigenetic clocks were observed after medications that improved metabolic health.

This included semaglutide used in a specific patient group with metabolic dysfunction and abnormal fat distribution.

Several clocks improved after treatment, while DunedinPACE suggested a slower pace of aging.

Potential mechanisms included:

Weight loss.
Reductions in visceral fat.
Improved glucose control.
Reduced insulin resistance.
Changes in protein and metabolite profiles.
Lower chronic inflammation.

These findings do not prove that semaglutide is a universal rejuvenation drug for healthy, lean individuals.

The trial involved a specific clinical population with identifiable metabolic abnormalities. The size of the improvement may have been partly related to the unfavorable starting point.

Results involving statins require similar caution. Some studies associated statin use with favorable changes in epigenetic aging clocks.

Statins reduce atherogenic lipoproteins and cardiovascular-event risk in appropriately selected patients. They may also influence inflammation and vascular function.

They should not, however, be prescribed solely to improve a biological-age test. Treatment decisions should be based on actual cardiovascular risk, ApoB or LDL concentration, coexisting conditions, and treatment tolerance.

The evidence for metformin was inconsistent. One small study reported improvements in selected clocks, while other experiments found no significant effect.

The available evidence therefore does not justify metformin use by healthy individuals solely to reduce epigenetic age.

The main value of these medications may lie in treating an existing disorder rather than directly rejuvenating the body.

Experimental therapies

The reviewed studies also included more experimental interventions, such as umbilical cord plasma products, plasma-exchange procedures, psychiatric medications, and other clinical therapies.

Some lowered selected epigenetic clock results, but the studies frequently involved very small groups.

This creates a high risk of overinterpretation. When a study includes only a small number of participants and analyzes many clocks, a positive result may arise by chance.

Other limitations include:

Lack of independent replication.
Absence of long-term follow-up.
High treatment costs.
Uncertain risks.
Selective reporting of positive findings.
No proof of reduced disease incidence or mortality.

Plasma-exchange findings were particularly interesting. In some analyses, the intervention was followed by a small increase rather than a decrease in epigenetic age.

This does not necessarily mean that plasma exchange permanently accelerates aging. The procedure may cause a temporary physiological response or alter the composition of blood cells.

It does show, however, that an attractive biological theory does not guarantee a favorable result in humans.

Experimental interventions should be evaluated much more rigorously than foundational lifestyle measures. The greater the cost, invasiveness, and potential risk, the stronger the evidence of real benefit should be.

What did not reduce epigenetic age?

A substantial proportion of the reviewed studies found no meaningful improvement in epigenetic aging clocks.

Interventions without a clear effect included:

Nicotinamide riboside.
Selected senolytics.
Short-term rapamycin use.
Some metformin studies.
Folic acid.
Polyphenol concentrates.
Certain weight-loss diets.
Selected exercise programs.
Several psychosocial interventions.
Some multi-ingredient supplements.

This does not automatically mean that all these interventions are completely ineffective.

A study may have been too short, the sample too small, or the selected clock insufficiently sensitive to the biological mechanism involved.

At the same time, an intervention should not be assumed to work merely because it is biologically interesting or effective in mice.

A lack of positive human results should reduce confidence, especially when the intervention has potential adverse effects.

Rapamycin is a useful example. It extends lifespan in several animal models, but a short human study did not produce clear improvements in the analyzed epigenetic clocks.

This does not settle the question of rapamycin’s effect on human aging, but it also does not confirm efficacy.

The same caution applies to NAD+ precursors. Their cellular mechanisms are interesting, but existing trials have not demonstrated consistent reductions in epigenetic age.

A plausible mechanism is not the same as a proven clinical effect.

Why can clock results be misleading?

An epigenetic clock is a statistical model, not a direct measure of the total wear and tear of the human body.

Its result can change for reasons that do not represent lasting rejuvenation.

Relevant factors include:

Timing of sample collection. Some epigenetic processes and immune-cell proportions vary during the day.
Acute infection. Immune activation may temporarily alter blood composition and methylation patterns.
Strenuous exercise. Intense training can temporarily change inflammation, stress hormones, and immune-cell counts.
Sleep deprivation and stress. Short-term physiological strain may worsen the result without indicating permanent acceleration of aging.
Laboratory variability. Sample collection, storage, and processing can introduce additional variation.
Regression to the mean. An unusually high first result may naturally move closer to the average during a later measurement.
Changes in blood-cell composition. An intervention can alter the proportions of lymphocytes, neutrophils, and other cells, affecting the clock without rejuvenating every tissue.

Another important issue is the analysis of many clocks at once. If researchers calculate ten separate results, there is a chance that one will improve randomly.

The most favorable result should therefore not be considered in isolation. It is important to examine whether:

The effect appeared across several clocks.
The difference was large enough to matter.
Multiple comparisons were taken into account.
The result differed from the control group.
Established health markers also improved.
The effect persisted over time.

A single reduction of two years does not necessarily mean that the body has been rejuvenated by two years. Part of the change may reflect natural measurement variability.

How to build a practical intervention hierarchy

Longevity interventions should not be selected solely according to the size of a change in an epigenetic clock.

A better approach combines four criteria:

The strength of evidence involving epigenetic age.
The effect on real disease risk.
Safety and long-term sustainability.
Relevance to the individual’s actual health limitations.

This produces a more practical hierarchy.

Level one: foundations with the highest value

The first priority should be interventions supported by broad evidence regardless of their effect on epigenetic clocks:

Regular endurance and resistance training.
Sufficient daily movement.
Avoiding smoking.
Maintaining healthy blood pressure.
Controlling ApoB, LDL, and glucose.
Obtaining sufficient sleep.
Preserving muscle mass.
Following a nutrient-dense diet.

These actions remain valuable even when an epigenetic test shows no improvement.

Level two: correcting genuine limitations

The next step is to identify factors that are limiting the health of a specific person:

Inadequate omega-3 intake.
Excess visceral fat.
Insulin resistance.
Vitamin or mineral deficiencies.
Poor cardiorespiratory fitness.
Chronic sleep deprivation.
Untreated hypertension.
Elevated cardiovascular risk.

Correcting one major problem will probably provide more benefit than adding another supplement to an already well-optimized routine.

Level three: careful personalization

At this stage, more specific strategies can be considered, such as adjusting calorie intake, measuring EPA and DHA status, refining the training program, or monitoring selected biomarkers.

The goal should not be to achieve the lowest possible epigenetic age at any cost. The more important objective is to build a system that can be sustained for years without reducing quality of life.

Level four: experimental interventions

Medications and invasive therapies should be used when there is a clear medical indication or within well-controlled clinical research.

An epigenetic clock result should not independently justify the use of rapamycin, metformin, semaglutide, senolytics, or expensive plasma-exchange procedures.

A potential biomarker improvement must be weighed against:

Adverse-effect risk.
Lack of long-term evidence.
Effects on immunity and metabolism.
Financial cost.
The possibility of worsening other health outcomes.

Longevity overkill begins when optimizing one experimental biomarker starts to damage the basic foundations of health.

Summary

Research shows that epigenetic age and the pace of aging measured by next-generation clocks can change in response to interventions.

The most promising findings involved:

Long-term reduction of excess calorie intake.
Diets rich in minimally processed plant foods.
Regular physical activity.
Omega-3 supplementation.
Combined omega-3, vitamin D, and exercise programs.
Effective treatment of metabolic disorders.

This does not mean that any of these methods has been definitively proven to reverse biological age.

In many studies, only one of several clocks improved. Effects were often small, participant groups were limited, and follow-up periods were short.

Several popular interventions, including NAD+ precursors, senolytics, short-term rapamycin use, and some metformin trials, did not produce consistent results.

The most important practical conclusion is that the most valuable interventions are those that improve both epigenetic clocks and well-established health outcomes.

Regular exercise, a high-quality diet, metabolic health, muscle preservation, adequate sleep, and treatment of genuine risk factors should take priority over experimental therapies.

An epigenetic clock can be an interesting monitoring tool. It should not replace measurements of fitness, strength, blood pressure, glucose regulation, blood lipids, body composition, or daily function.

Lowering a test result is not the ultimate objective. The real goal is a longer life with better health, physical capacity, and independence.