Can Your Diet Really Slow Biological Ageing? What the Latest Epigenetic Research Actually Shows

The idea that we possess two different ages has become one of the most compelling concepts in longevity science. There is our chronological age, determined simply by the number of years we have been alive, and then there is our supposed biological age, an attempt to quantify how rapidly the tissues and systems within our body are actually ageing.

This distinction is intuitively attractive because chronological age clearly does not tell us everything about health. Two people can both be 65 years old while having profoundly different levels of cardiovascular fitness, metabolic health, muscle strength, cognitive function and chronic disease. One may remain physically capable and metabolically healthy while another has already accumulated several conditions normally associated with later life. It therefore seems entirely reasonable to ask whether we can measure some aspect of this difference biologically.

Epigenetic clocks have become one of the most sophisticated attempts to do exactly that. By analysing patterns of DNA methylation, researchers have developed algorithms capable of estimating chronological age and, increasingly, predicting aspects of disease, mortality and physiological decline. This has inevitably created another fascinating question. If lifestyle affects the biological processes involved in ageing, could changing the way we eat actually slow these clocks?

Two important studies published within days of each other in September 2026 illustrate both the enormous potential of this field and the considerable difficulty involved in interpreting it. The first developed a dietary pattern specifically associated with slower epigenetic ageing and found striking associations with biological-age measurements, mortality, chronic disease and physical function. The second used genetically identical twins to investigate whether the apparent relationship between a healthy diet and slower epigenetic ageing remained when genetics and shared family environment were taken into account.

Together, the studies tell a much more interesting story than either would in isolation. Diet quality is consistently associated with healthier ageing, and epigenetic clocks may eventually help us understand some of the biological processes involved. What remains much less certain is whether eating a particular diet directly turns back an ageing clock, or whether these measurements are partly reflecting the same complex collection of genetic, environmental and behavioural characteristics that already distinguish healthier people from less healthy people.

Understanding that distinction is essential before biological age becomes another nutritional number that people attempt to optimise without understanding what it actually represents.

What is biological age?

Chronological age is straightforward. Biological age is not.

There is no single universally accepted biological-age measurement in the way that there is a standard method for measuring blood pressure or blood glucose. Instead, researchers have developed numerous methods intended to capture different features associated with ageing. These include combinations of clinical biomarkers, measures of physical function, inflammatory markers, proteomic and metabolomic profiles and, increasingly, patterns of epigenetic change.

Epigenetic clocks are among the best-known of these approaches. They examine chemical modifications to DNA, particularly DNA methylation, which involves the addition of methyl groups at specific sites within the genome. These modifications do not change the underlying DNA sequence, but they can be associated with regulation of gene activity and numerous biological processes.

Patterns of DNA methylation change in remarkably predictable ways as humans grow older. Researchers discovered that by measuring methylation at particular sites and combining these measurements using statistical algorithms, it was possible to estimate chronological age with surprising accuracy.

The first generation of epigenetic clocks was largely designed to predict chronological age. Later generations became more ambitious. Rather than simply asking how old someone was, researchers began constructing clocks using biological and clinical information associated with morbidity and mortality. These newer clocks attempt to capture aspects of ageing that may be more relevant to health outcomes.

This distinction is important because the phrase “epigenetic age” can create the impression that researchers have discovered a single molecular speedometer showing exactly how quickly somebody is ageing. In reality, different clocks have been developed using different datasets, different statistical methods and different outcomes. They consequently measure overlapping but not identical aspects of biology.

GrimAge and DunedinPACE measure different aspects of ageing

Two measurements appear repeatedly in recent nutrition research: GrimAge and DunedinPACE.

GrimAge was developed partly using DNA methylation surrogates for plasma proteins and smoking exposure that are associated with morbidity and mortality. More recent versions, including GrimAge2, incorporate additional information intended to improve prediction of health outcomes. A person whose GrimAge is greater than would be expected from their chronological age may be described as showing epigenetic age acceleration.

DunedinPACE takes a somewhat different approach. It was developed using longitudinal information from the Dunedin Study and attempts to estimate the pace at which ageing-related physiological changes are occurring. Rather than simply producing an estimated biological age in years, it aims to capture whether someone appears to be ageing physiologically more quickly or slowly.

These measures are scientifically valuable because accelerated epigenetic ageing has been associated with numerous adverse outcomes, including chronic disease and mortality. That makes them potentially useful research tools for investigating why some people age more successfully than others and whether interventions can modify ageing-related biology.

However, an association with mortality does not automatically mean that the clock itself is a causal mechanism of ageing. A speedometer tells us how quickly a car is travelling, but changing the number displayed on the speedometer does not necessarily change the speed of the car. Similarly, an ageing biomarker can be extremely good at identifying people with greater disease risk without necessarily being something that needs to be directly manipulated.

This distinction becomes particularly important when researchers begin using these clocks to evaluate diet.

A new dietary pattern was associated with substantially slower epigenetic ageing

A major study published in Nature Communications in September 2026 approached the relationship between food and epigenetic ageing in an interesting way. Rather than beginning with an established dietary pattern such as the Mediterranean diet or DASH diet, researchers developed an empirical dietary index specifically designed to identify dietary characteristics associated with slower epigenetic ageing.

The researchers called this the Empirical Dietary Index for Slower Epigenetic Aging, or EDISEA. The index was developed using GrimAge2 acceleration as the biological-age outcome, allowing dietary information to be used to identify a food pattern associated with more favourable epigenetic ageing.

In the initial sample of 1,711 participants, greater adherence to the EDISEA pattern was associated with approximately 4.17 years slower epigenetic ageing. Importantly, the researchers then tested the dietary index in an independent sample of 2,311 people, where greater adherence was again associated with slower epigenetic ageing, this time by approximately 3.77 years.

Those are striking differences and understandably make for attractive headlines. A dietary pattern associated with being almost four years biologically younger appears to provide exactly the kind of evidence people interested in longevity have been waiting for.

The researchers did not stop with the epigenetic clock. Higher EDISEA scores were also associated with a lower risk of all-cause mortality across two US cohorts, with the pooled analysis suggesting approximately a 20% lower risk among those with higher scores. Greater adherence was additionally associated with lower risks of several ageing-related diseases, functional limitations and multimorbidity across US and UK ageing cohorts.

This broader relationship with actual health outcomes makes the study considerably more interesting than one showing only a change in an epigenetic biomarker. If a dietary pattern associated with slower epigenetic ageing is also associated with lower mortality, fewer diseases and better physical function, it raises the possibility that the clock may be detecting biologically meaningful differences between people.

However, it still does not establish that the diet itself has turned back biological ageing.

Association is not the same thing as reversing ageing

This is where biological-age research can become particularly vulnerable to exaggeration.

Suppose researchers identify people eating healthier diets and find that those people have lower GrimAge scores. There are several possible explanations. The diet may genuinely be influencing biological processes that alter DNA methylation and contribute to healthier ageing. That is entirely plausible.

But people who eat healthier diets frequently differ from people eating poorer diets in numerous other ways. They may exercise more, smoke less, have different levels of adiposity, consume less alcohol, have higher socioeconomic status, sleep differently, use healthcare differently or have other behaviours that influence long-term health.

Researchers can statistically adjust for many of these factors, but statistical adjustment is never perfect. Some variables are measured imprecisely, some are unknown and others may interact in ways that are difficult to model.

There is also the possibility that some of the association reflects genetics. Genetic variation can influence food preference, appetite, metabolism, body composition, disease susceptibility and numerous behavioural characteristics. If some of those same genetic influences are also associated with epigenetic ageing, diet and biological age may appear more directly connected than they really are.

This is why observational nutrition research can identify important patterns without necessarily establishing causality. The EDISEA findings are compelling, particularly because associations were reproduced across different populations and extended beyond an epigenetic clock to disease and mortality outcomes. Nevertheless, the study cannot tell us that adopting this particular dietary pattern will make an individual four years biologically younger.

Remarkably, another study published the following day provided an almost perfect demonstration of why that caution is necessary.

Identical twins allowed researchers to investigate the role of genetics

A study published in GeroScience used data from the Danish GEMINAKAR cohort to investigate the relationship between diet quality and epigenetic ageing in monozygotic twins.

Identical twins provide a particularly powerful natural experiment because members of a monozygotic twin pair share essentially the same genetic sequence and much of their early-life environment. Comparing unrelated people can leave substantial differences in genetics and upbringing between individuals. Comparing one identical twin with the other allows researchers to control much more effectively for these familial influences.

The study included 132 pairs of monozygotic twins with information collected at baseline and again approximately 12 years later. Dietary quality was assessed according to adherence to Danish dietary guidelines, while epigenetic ageing was measured using several established clocks, including principal-component versions of Hannum, Horvath, PhenoAge and GrimAge, alongside DunedinPACE.

When researchers initially analysed participants as individuals, the results looked reassuringly familiar. Better adherence to dietary guidelines was associated with slower epigenetic ageing, particularly when assessed using PCGrimAge and DunedinPACE.

For every one-unit increase in the dietary-adherence score, PCGrimAge was approximately 0.55 to 0.67 years lower, while DunedinPACE suggested approximately a 3.1% slower pace of ageing.

If the analysis had ended there, the study could easily have joined the growing collection of papers supporting the claim that healthier diets slow biological ageing.

It did not end there.

The relationship weakened when one identical twin was compared with the other

The researchers then performed within-twin-pair analyses. Instead of asking whether people with healthier diets generally had younger epigenetic ages, they effectively asked whether the twin who adhered more closely to healthy dietary guidelines also showed meaningfully slower epigenetic ageing than their genetically identical sibling.

The associations became substantially weaker.

The particularly noticeable relationships observed for PCGrimAge and DunedinPACE in the individual-level analyses were attenuated when comparisons were made within twin pairs. The researchers also failed to identify strong evidence from their longitudinal analyses that improvements in dietary adherence over the approximately 12-year follow-up were associated with corresponding improvements in epigenetic ageing.

This suggests that at least part of the apparent relationship between diet quality and epigenetic ageing may be explained by genetic factors or aspects of the shared family environment rather than a simple direct effect of diet itself.

It is important not to overinterpret this study in the opposite direction. The sample was relatively small, particularly once analyses were conducted within twin pairs, which reduces statistical power. Dietary intake was assessed using food-frequency questionnaires, which inevitably introduce measurement error. The researchers also could not distinguish precisely how much of the attenuation was attributable to genetics and how much reflected shared environmental influences.

Nevertheless, the design addresses a limitation that conventional observational studies struggle to overcome. It demonstrates that an apparently persuasive association between healthy eating and slower epigenetic ageing can become considerably less convincing when familial confounding is controlled more rigorously.

That does not prove diet has no effect on epigenetic ageing. It tells us that the relationship is more complicated than a simple equation in which healthier food automatically produces a younger methylation clock.

Other recent research still finds relationships between diet quality and epigenetic ageing

The twin findings also need to be considered within the wider literature rather than treated as the final answer.

Another large study published in Nature Communications in August 2026 examined ten different diet-quality scores using two population-based cohorts. Researchers found surprisingly little overlap between the people classified as having the highest adherence to different healthy dietary patterns, illustrating how differently various dietary indices can define a “healthy diet”.

Despite those differences, greater adherence to healthy dietary patterns was generally associated with reduced epigenetic ageing. Different dietary patterns were also associated with distinct DNA methylation signatures, but many of these appeared to converge on similar biological pathways.

This is potentially important because it echoes a broader finding within nutrition research. There may not be one uniquely optimal dietary pattern for healthy ageing. Mediterranean-style diets, DASH, MIND and other high-quality dietary patterns differ in their exact composition, yet frequently share several characteristics including greater consumption of vegetables, fruits, legumes, nuts and whole grains and lower reliance on heavily refined foods and processed meats.

If multiple dietary patterns produce similar associations with ageing-related biology, the common characteristics may matter more than allegiance to one named diet.

However, the twin study reminds us that even consistent associations across different dietary indices do not automatically prove that the observed epigenetic differences are caused directly by diet.

Can food actually alter DNA methylation?

There is no biological reason to assume that diet cannot influence epigenetic regulation. Nutritional status can affect numerous biochemical processes involved in methyl-group metabolism, cellular signalling, inflammation, oxidative stress and energy metabolism, all of which can interact with epigenetic regulation.

Nutrients involved in one-carbon metabolism provide an obvious example. Folate, vitamin B12, vitamin B6, choline and methionine participate in pathways involved in methyl-group availability. Numerous phytochemicals and metabolic signals can also interact with enzymes and signalling pathways involved in gene regulation.

Diet additionally influences metabolic health, body composition, inflammatory signalling and the gut microbiome, each of which could potentially affect epigenetic patterns indirectly.

The mistake would therefore be to move from the reasonable statement that diet can influence epigenetic biology to the much stronger claim that eating a particular collection of foods has been demonstrated to reverse human ageing.

DNA methylation is dynamic and responsive to environmental influences, but it is also tissue-specific and enormously complex. Most large human studies measure methylation in blood because blood is relatively easy to obtain. An epigenetic-age measurement derived from circulating blood cells does not necessarily tell us precisely what is happening simultaneously in skeletal muscle, brain tissue, liver, heart or other organs.

This does not make blood-based epigenetic clocks useless. It simply defines what they are: sophisticated biomarkers constructed from measurable biological information rather than direct measurements of the entire ageing process.

A younger epigenetic age may be useful without being the goal itself

One of the most important questions in this field is whether epigenetic clocks are merely markers of ageing or whether changing them represents a meaningful therapeutic objective.

If people with accelerated GrimAge consistently experience more disease and earlier mortality, the measurement clearly contains useful information. It may help researchers identify biological pathways associated with unhealthy ageing, evaluate populations at greater risk or investigate whether interventions influence ageing-related biology.

However, the value of a biomarker as a predictor does not automatically make the biomarker itself the thing we need to optimise.

This distinction is familiar throughout medicine. Numerous biomarkers correlate with disease risk because they reflect underlying physiology. Some are themselves causal contributors to disease and become useful therapeutic targets. Others primarily tell us something about the state of the system without necessarily being appropriate targets in their own right.

Epigenetic clocks may contain elements of both. Some methylation changes may participate directly in biological processes relevant to ageing, while others may primarily record the accumulated consequences of smoking, metabolic dysfunction, inflammation, environmental exposure or other influences.

This is one reason intervention studies are so important. If a dietary intervention changes an epigenetic clock, researchers then need to establish whether that change corresponds with improvements in outcomes that matter, including metabolic health, cardiovascular disease, physical function, cognition, frailty and ultimately mortality.

A clock becoming younger is interesting. A person remaining healthier and capable for longer is the objective.

The commercial biological-age industry deserves particular caution

As epigenetic clocks have moved beyond research laboratories, direct-to-consumer biological-age testing has expanded rapidly. People can now send samples to commercial laboratories and receive a biological-age estimate that may tell them they are several years younger or older than their chronological age.

The attraction is obvious. A single number appears to summarise the enormously complicated process of ageing, and repeating the test after changing diet, exercise or supplements creates the possibility of watching one's biological age apparently move backwards.

The scientific interpretation is considerably more difficult.

Different clocks can produce different estimates because they were trained using different outcomes and different methylation sites. Biological samples, laboratory procedures, statistical methods and normal measurement variation can all influence results. Researchers are also still determining how much change in a particular clock represents a clinically meaningful alteration in long-term health.

A person whose reported biological age falls from 58 to 54 after changing their diet may understandably conclude that they have reversed four years of ageing. That conclusion requires considerably more evidence than the number alone provides.

The test may have detected a real change in DNA methylation. The change may even reflect improvements in aspects of health. What cannot automatically be concluded is that every organ in the person's body has become four years younger or that four years have been added to their lifespan.

Biological-age measurements should therefore be regarded as emerging research tools rather than definitive report cards on how successfully an individual is ageing.

None of this weakens the evidence for eating well

There is a danger when discussing uncertainty around biomarkers that the broader message becomes distorted. If the relationship between healthy eating and epigenetic ageing is partly confounded, that does not mean diet suddenly becomes irrelevant to healthy ageing.

We do not need epigenetic clocks to establish that nutrition affects cardiovascular and metabolic health. Decades of research already demonstrate relationships between dietary patterns and blood pressure, blood lipids, glucose regulation, body composition, cardiovascular disease, type 2 diabetes and numerous other outcomes relevant to healthspan.

Long-term prospective research has also associated high-quality dietary patterns with a greater likelihood of reaching older age free from major chronic disease while retaining physical, cognitive and mental health. These outcomes are considerably more tangible than a biological-age score because they describe whether people actually remain healthier as they grow older.

This distinction matters because there is a tendency within longevity culture to believe that newer molecular measurements somehow supersede conventional health outcomes. They do not. An epigenetic clock can provide fascinating additional information about ageing biology, but it does not make blood pressure, LDL cholesterol, insulin sensitivity, muscle strength, cardiorespiratory fitness or physical function obsolete.

If somebody improves their diet, reduces excess visceral fat, improves their blood pressure, increases their fitness and reduces their risk of cardiovascular disease, those changes remain profoundly valuable regardless of whether a methylation algorithm reports that they are three years biologically younger.

What should a diet for healthy ageing actually look like?

The interesting feature of the emerging epigenetic research is that it does not appear to be revealing an entirely new collection of longevity foods. The dietary patterns associated with healthier epigenetic profiles broadly overlap with dietary characteristics already associated with cardiovascular, metabolic and overall health.

This means an eating pattern centred around a wide variety of vegetables and fruits, legumes, nuts, seeds and other minimally processed plant foods remains a sensible foundation. Whole grains can contribute fibre and micronutrients where they are tolerated and preferred, while unsaturated fats from foods such as nuts, seeds, olive oil, avocado and oily fish can form the predominant sources of dietary fat.

Protein also becomes increasingly important as people age because preserving muscle mass and function is fundamental to maintaining independence and metabolic health. That protein can come from a combination of plant and animal foods according to dietary preference, but the overall intake needs to be sufficient to support muscle alongside resistance exercise.

The dietary pattern should also support appropriate energy balance and body composition. Excess visceral adiposity, insulin resistance, hypertension and dyslipidaemia are established threats to healthy ageing, and there is little sense in becoming preoccupied with epigenetic optimisation while ignoring them.

Similarly, highly refined and ultra-processed foods do not need to be treated as poisons, but a dietary pattern dominated by them can displace foods providing fibre, micronutrients, unsaturated fats, protein and intact food structure. The overall pattern across years matters considerably more than achieving dietary perfection at every meal.

Perhaps the most reassuring message is that healthy-ageing nutrition remains remarkably recognisable. The frontier science may involve DNA methylation and machine-learning algorithms, but the foods repeatedly associated with healthier outcomes remain remarkably conventional.

We need trials that change the diet and follow what happens next

The next important stage in this research is not simply to produce more observational studies showing that people with healthier diets have younger epigenetic ages. We already have substantial evidence of that association.

What we need are well-designed randomised dietary intervention trials that deliberately alter dietary patterns and then measure how epigenetic clocks respond alongside established clinical and functional outcomes. These studies need adequate sample sizes and sufficiently long follow-up to distinguish genuine biological changes from short-term variation.

Ideally, researchers would measure multiple epigenetic clocks rather than selecting one favourable result and would investigate whether changes in biological-age measurements correspond with improvements in cardiovascular risk, metabolic health, cognition, muscle function and other meaningful outcomes. Longer follow-up could then determine whether participants experiencing favourable changes in epigenetic ageing actually develop less disease or maintain better function.

This is a much higher evidential standard than simply observing that people eating more vegetables have younger GrimAge scores, but it is precisely the standard required if epigenetic clocks are going to move from fascinating research instruments to validated tools for evaluating longevity interventions.

The identical-twin research adds another important requirement. Future studies need to continue addressing genetic and familial confounding rather than assuming that associations remaining after conventional statistical adjustment are necessarily causal.

So, can diet slow biological ageing?

The most scientifically defensible answer at present is that diet almost certainly influences many of the biological processes involved in ageing, and healthier dietary patterns are repeatedly associated with more favourable epigenetic-age measurements. The new EDISEA research strengthens that observation by showing that a dietary pattern derived specifically from slower GrimAge2 ageing was also associated with lower mortality, fewer ageing-related diseases and better functional outcomes across several cohorts.

Those findings are important and deserve attention. They suggest that epigenetic clocks may be capturing biologically meaningful differences associated with dietary patterns and long-term health.

The new identical-twin study adds an equally important qualification. When researchers controlled more rigorously for genetics and shared family environment, the apparent relationship between diet quality and epigenetic ageing became considerably weaker, while longitudinal analyses did not provide strong evidence that improvements in dietary adherence produced corresponding reductions in epigenetic ageing.

These findings are not contradictory. They represent different pieces of the same scientific problem. One body of evidence tells us that healthy dietary patterns and slower epigenetic ageing frequently occur together. The other warns us that we have not yet established how much of that relationship is directly caused by diet.

This is exactly how nutrition science should progress. An exciting association is identified, increasingly sophisticated studies test alternative explanations, and our interpretation becomes more precise.

Healthy ageing matters more than achieving a younger number

The growing ability to quantify ageing biology is one of the most exciting developments in modern geroscience. Epigenetic clocks may eventually help researchers identify people ageing unusually quickly, understand the biological consequences of lifestyle exposures and test interventions designed to extend healthspan. As the technology improves, these measurements may become genuinely useful components of preventive medicine.

For now, however, they should not distract us from the fundamental objective. The goal of healthy-ageing nutrition is not to persuade an algorithm that we are younger than the date on our birth certificate. It is to reduce the probability of developing the diseases and functional limitations that make ageing less healthy.

A dietary pattern that supports cardiovascular health, maintains glucose regulation, provides sufficient protein and micronutrients, supports skeletal muscle and bone, promotes an appropriate body composition and supplies plenty of fibre and minimally processed plant foods already addresses many of the processes that determine whether we remain capable as we grow older.

If future research demonstrates that such a diet also reliably slows validated measures of biological ageing, that will add an important new layer to our understanding. It would help explain some of the molecular pathways through which healthy dietary patterns exert their effects and may eventually allow interventions to become more personalised.

Until then, a biological-age score should remain what it currently is: an intriguing and increasingly sophisticated window into ageing biology rather than a definitive measurement of how many years we have gained or lost.

The strongest reason to eat well remains considerably more tangible. We want to reach older age with a cardiovascular system, metabolism, brain, muscles and bones that still allow us to live independently and participate fully in life. Whether an epigenetic clock eventually describes that state as three years, five years or ten years younger is scientifically fascinating, but it is not the outcome that matters most.

References

Lai, S., Yu, J., Zhang, L. et al. (2026) ‘An epigenetic aging-informed dietary pattern is associated with a spectrum of aging-related health outcomes’, Nature Communications. doi: 10.1038/s41467-026-77790-9.

Nygaard, M., Elliott, H.R., Davey Smith, G., Relton, C., Kyvik, K.O. et al. (2026) ‘Dietary pattern and epigenetic aging: a longitudinal twin study’, GeroScience. doi: 10.1007/s11357-026-02550-y.

Tavares, J.F., Liu, D., Talevi, V., Eichelmann, F., Jannasch, F., Schulze, M.B., Aziz, N.A., Nöthlings, U. et al. (2026) ‘Associations between diet quality, epigenetic aging and epigenome in two population-based cohorts’, Nature Communications, 17, 9232. doi: 10.1038/s41467-026-77064-4.

Teschendorff, A.E. and Horvath, S. (2025) ‘Epigenetic ageing clocks: statistical methods and emerging computational challenges’, Nature Reviews Genetics, 26(5), pp. 350–368. doi: 10.1038/s41576-024-00807-w.

Fadadu, R.P. et al. (2025) ‘Chemical and climatic environmental exposures and epigenetic aging: a systematic review’, Environmental Research, 274, 121347. doi: 10.1016/j.envres.2025.121347.