Biological age: what is it and can it be measured?

Biological age is an attempt to assess how closely the body's condition corresponds to the typical condition of people of a given calendar age. However, it is not a single, directly measurable indicator, like temperature, blood pressure, or glucose levels. Today, dozens of mathematical models exist that use blood tests, DNA methylation, plasma proteins, functional indicators, organ imaging, and other data. There is currently no single "gold standard" for biological age. [1]

Therefore, the phrase "I'm 45 years old, but my biological age is 37" requires clarification: how was this 37-year-old figure obtained? The Horvath epigenetic clock, GrimAge, PhenoAge, DunedinPACE, the model based on routine blood tests, or the plasma protein algorithm all measure different things. A 2026 study comparing five widely used epigenetic clocks found that they are associated with different molecular processes and contain more unique than common biological signals. [2]

Modern biological clocks are very useful in aging research and are already capable of providing additional information about disease risk, functional decline, and mortality at the population level. However, a commercial test should not yet be interpreted as a medical diagnosis, a precise estimate of "years of life remaining," or proof that the body has truly become ten years younger. The international Biomarkers of Aging Consortium considers the clinical translation of such indicators promising, but emphasizes the need for better individual reproducibility, clinical interpretability, and linking results to specific medical interventions. [3]

What is biological age?

Chronological age only answers the question of how much time has passed since birth. It increases equally for everyone: every year, everyone becomes a year older.

However, physiologically, people age differently. Two 60-year-olds can have significantly different blood pressure, vascular health, kidney function, muscle strength, physical endurance, glucose metabolism, inflammatory markers, and cognitive abilities. It is this heterogeneity that the concept of biological age attempts to quantify. [4]

The difficulty lies in the fact that aging occurs simultaneously on many levels. The functioning of cells, the immune system, blood vessels, muscles, the brain, kidneys, and other organs changes; molecular damage accumulates; gene expression and epigenetic regulation change. Different processes can proceed at different rates even within a single organism. Therefore, attempting to reduce all of human aging to a single number is inevitably a model, not a direct measurement of some hidden physical quantity. [5]

For this reason, the World Health Organization's concept of healthy aging focuses not on calculating the "true age of the body," but on a person's intrinsic potential and functional capacity—the ability to move, think, see, hear, make decisions, and maintain independence. This approach focuses directly on a person's health and function. [6]

Is it really possible to measure biological age?

It is possible to obtain a scientifically based assessment of biological aging, but it is not possible today to measure a single objective “true biological age.”

A systematic review of methods for estimating biological age found a large number of approaches and statistical models, but the authors concluded that there is still no agreed-upon gold standard. The results depend on which biomarkers are included in the model and the population on which it was developed. [7]

This is fundamentally different from measuring, for example, glucose concentration. If a lab detects glucose at 5.4 mmol/L, it's referring to a physically measurable concentration of the substance. If an algorithm returns a "biological age of 52," it's effectively saying that, according to the rules of this model, your set of indicators is more similar to a certain reference profile corresponding roughly to that age or risk level.

Therefore, biological age is more correctly viewed as a calculated index of condition or risk, rather than as a second date of birth.

Why two tests can show different biological ages

Because they can measure different aspects of aging, this doesn't necessarily mean one of the tests is wrong.

In 2026, researchers simultaneously analyzed five widely used epigenetic clocks—Horvath, Hannum, PhenoAge, GrimAge, and DunedinPACE—in 3,227 participants in the Health and Retirement Study. Comparisons with gene activity revealed marked heterogeneity: each algorithm was associated with its own set of molecular pathways, and unique biological signals outnumbered shared ones. [8]

A similar problem exists with models based on standard clinical biochemistry. Comparisons of several such indicators across different cohort studies showed that different metrics of biological aging correlated rather weakly with each other and could predict individual outcomes differently. [9]

Therefore, it is quite possible that a person will have:

• Horvath watches show age close to passport age;
• GrimAge estimates the risk as slightly higher;
• PhenoAge shows "younger";
• DunedinPACE indicates a normal or increased current rate of ageing.

These numbers cannot simply be averaged and the average declared to be the “true biological age.”

Basic methods for measuring biological age

Method	What is it based on?	Which roughly reflects	The main problem
Clinical hours	routine blood tests and physiological indicators	the state of several systems and the risk of adverse outcomes	depends on the selected indicators and population
Epigenetic clock	DNA methylation	age-related molecular changes and/or risks	different clocks measure different signals
DunedinPACE	DNA methylation	estimated rate of aging	this is not age in the usual sense
Proteomic clock	hundreds or thousands of blood proteins	systemic and sometimes organ-specific aging	so far mainly a research tool
Telomeres	telomere length of cells	one of the processes associated with aging	too variable for a full "age of the organism"
Functional indicators	strength, mobility, respiratory function, etc.	functional aging	do not give a single molecular age

Each approach may answer a useful question, but none yet provides a universal standard. [10]

Biological age based on routine blood tests

Aging can be assessed even without DNA analysis. Some algorithms use parameters already measured in standard clinical practice.

One of the best-known examples is Phenotypic Age. Its basic model combines calendar age with nine laboratory parameters: albumin, alkaline phosphatase, creatinine, C-reactive protein, glucose, white blood cell count, lymphocyte percentage, mean corpuscular volume, and red blood cell distribution width. A mathematical model then converts the resulting risk into a value expressed in years. [11]

This makes these watches attractive: many of the necessary values are already present in a standard blood and biochemistry test. Costly genomic analysis is not required.

However, accessibility does not mean absolute accuracy. A 2026 systematic review identified 81 aging models based on routine clinical indicators and revealed significant issues with the quality of development and external validation of many algorithms. A significant portion of the models were developed using only the population of a single country, and more than a third lacked either internal or external validation. [12]

Therefore, a free calculator that asks you to enter several blood tests does not automatically become a medically validated test just because it uses real lab results.

What does Phenotypic Age mean?

Phenotypic Age is a good example of why the term "biological age" is sometimes misleading.

This model was not actually built to find some internal counter for the number of years lived by cells. It combines laboratory parameters related to health and mortality, and then expresses the resulting risk in familiar units of age. [13]

Therefore, if a 50-year-old person receives a Phenotypic Age of 58 years, it is more correct to interpret this as:

"Based on the combination of biomarkers used by the model, your profile corresponds to a higher level of risk than is typically expected for your chronological age."

It is much less correct to say:

"Your body is literally 58 years old."

This is why the idea of using the term risk-equivalent age has emerged in scientific literature, as it better describes the meaning of some modern clocks. [14]

Epigenetic age: why is it talked about most often?

Epigenetic clocks are one of the most studied classes of aging biomarkers. They analyze DNA methylation: chemical markers on specific regions of the genome that change regularly with age.

Methylation does not change the DNA sequence. It is part of epigenetic regulation—a system of chemical changes that influence the use of genetic information.

In 2013, Steve Horvath created a famous multi-tissue clock using methylation of 353 CpG DNA regions. The algorithm was developed on thousands of tissue samples and was effective in estimating chronological age. This work largely made the term "epigenetic clock" widely known. [15]

But the ability to accurately predict age from DNA does not mean the model automatically measures health or the rate of aging. First-generation watches, including Horvath and Hannum, were primarily trained to predict chronological age. [16]

Later, models emerged that focused specifically on health, disease, and survival.

PhenoAge

DNAm PhenoAge was developed using a clinical phenotypic model of aging. In initial work, it predicted a range of outcomes better than early hours, including overall mortality, physical functioning, and several age-related diseases. [17]

That is, the PhenoAge's purpose is already different from that of the classic Horvath watch: the model attempts to capture not only the pattern of "how old the sample is," but also the differences in the health of people of the same age.

GrimAge

GrimAge was developed with an even stronger focus on prognosis. The model includes epigenetic surrogates of several plasma proteins and an index associated with cumulative tobacco exposure. In its initial validation, GrimAge was associated with mortality and a number of age-related outcomes. [18]

Therefore, "GrimAge 60" and "Horvath Age 60" are conceptually different results, although both are expressed in years.

A large-scale comparison of 14 epigenetic clocks in 2025 confirmed that second- and third-generation clocks perform, on average, better than earlier models in disease prediction tasks. However, even for the best models, the increase in accuracy over traditional risk factors was uneven across different diseases. [19]

DunedinPACE: It's not age, it's the speed of aging

DunedinPACE answers a different question: not "how old does an organism look?", but "how fast is biological aging currently occurring according to a given model?"

The method grew out of the Dunedin Study, a long-term study of people of the same chronological age. Researchers repeatedly measured 19 indicators of the health of various body systems over approximately two decades and calculated individual rates of change. This long-term physiological profile was then translated into an algorithm capable of assessing the rate of aging based on DNA methylation from a single blood sample. [20]

In this sense, DunedinPACE is closer to a "speedometer" than an odometer.

Conventionally, a result of about 1 corresponds to a rate of approximately one biological year of change per calendar year in the model's reference system; higher values are interpreted as a faster rate, lower values as a slower one. However, this is a statistical value, not a literal observation of how a person has "aged by 1.2 years" in a year. [21]

This distinction between age and rate is fundamentally useful. A person may already have an accumulated unfavorable physiological profile, but currently age relatively slowly. Conversely, their current "age" may be favorable, but the rate of change in some systems may be accelerated.

Age of individual organs: can the heart be “older” than the brain?

Modern research shows that aging can indeed be asynchronous: different organs and tissues within a single person change at different rates. Therefore, the idea of a single age for the entire organism may be too crude.

Stanford University researchers developed proteomic models—algorithms that use plasma proteins associated with specific organs. In a study published in Nature, they used this method to assess the aging of 11 major organs. Organ-specific acceleration was associated with diseases of the corresponding systems and subsequent risks. [22]

Another large proteomic algorithm was created using data from over 45,000 UK Biobank participants, measuring nearly 2,900 plasma proteins. The resulting proteomic age was associated with multiple chronic diseases, multimorbidity, physical and cognitive characteristics, and mortality; the model was further validated in populations from China and Finland. [23]

This is one of the most promising areas of geroscience. However, the results of such studies do not yet mean that healthy individuals have a standard medical indication for annually testing their "heart, brain, and liver age." Such models are still primarily under research and require further standardization and evaluation of individual clinical benefits. [24]

Is it possible to determine biological age by telomere length?

Telomere length is associated with aging, but it cannot be used as the only reliable test of whole-body age.

Telomeres are repetitive sections of DNA at the ends of chromosomes that protect them during cell division. In some cells, telomeres gradually shorten, so the idea of using their length as a kind of cellular counter arose logically.

But human biology turned out to be much more complex. A large systematic review and meta-analysis of 414 samples, including more than 743,000 people, found only a relatively weak mean association between telomere length and chronological age: the pooled correlation was approximately -0.19. The result depended significantly on the tissue and the laboratory measurement method. [25]

Different cells also behave differently. Postmitotic cells, which are virtually non-dividing, do not undergo the same mechanism of sequential telomere shortening as actively dividing cells. Telomeres are additionally affected by oxidative stress and other processes. [26]

Therefore, the result of the commercial test:

Telomeric age 42 years

It should not be taken as an exact measurement of the age of the entire organism.

Telomeres are one biologically interesting component of aging, but they are not a universal measure of how long the body has lived.

Is it possible to determine age from a blood test without a special test?

Routine laboratory parameters can indeed be combined mathematically to produce a research estimate of biological age. But creatinine, glucose, or C-reactive protein alone do not constitute a test for "aging."

These algorithms are attractive because they use readily available data. These may include indicators of kidney function, inflammation, glucose metabolism, blood cell composition, and other markers.

But the result is very sensitive to context.

For example, C-reactive protein temporarily increases during infection or inflammation. Glucose levels are dependent on nutrition and metabolic status. Creatinine is linked not only to kidney function but also to muscle mass. Therefore, a one-time unfavorable indicator can shift the calculated "age," even though a person has clearly not biologically aged by several years in a week.

This is why repeatability, robustness to short-term changes, and proven ability to predict truly important medical outcomes are essential for aging clocks. The Biomarkers of Aging Consortium considers individual reliability and the ability to respond to real-world health changes to be key requirements for further clinical implementation. [27]

Are fitness tests, grip strength and VO₂max biological age?

They may be important indicators of functional aging, but a single value like grip strength or maximal oxygen consumption is not a complete measure of the biological age of the entire organism.

Functional indicators have a significant practical advantage: they directly describe what a person is capable of doing. A decline in strength, mobility, or cognitive function is often much more clinically understandable than an abstract two-year change in epigenetic age.

That is why the World Health Organization's concept of healthy aging is based on preserving a person's inner potential and functional capacity. This concept includes mobility, the ability to think, see, hear, and maintain independence. [28]

This doesn't mean that molecular clocks are useless. Rather, it's about two levels of measurement: molecular methods can identify processes long before obvious functional decline, while functional indicators reveal their real impact on daily life.

How accurate is the epigenetic clock?

Here it is necessary to first define what exactly is meant by accuracy.

First-generation watches can predict chronological age very well. But if the algorithm predicts that a 50-year-old person's DNA is approximately 49-51 years old, this primarily demonstrates that the model is good at recognizing age-related epigenetic patterns. This does not prove that it perfectly measures the rate of aging or an individual's risk of death.

Modern second- and third-generation clocks are constructed differently and are better associated with outcomes such as disease, physical decline, and mortality. For example, in a study of 18,859 people, a comparison of 14 epigenetic clocks showed a significant advantage of newer generations over earlier age-predicting models for many diseases. [29]

But even this success relates primarily to population group statistics.

In 2025, a separate critical analysis of the individual use of epigenetic clocks reached a significantly more cautious conclusion: technical variability, sample processing characteristics, algorithm selection, tissue specificity, and biological fluctuations currently limit the ability to make decisions for a single person based on such a result. The authors explicitly consider current clocks insufficiently mature for individual clinical decisions. [30]

Therefore, the statement "an epigenetic test measures biological age with an accuracy of ±3 years" can be misleading. This formulation often describes the error in predicting passport age, rather than the accuracy of measuring a known true biological age—after all, there is no standard to compare it to.

What does "biological age younger than passport age" mean?

This is usually a good statistical signal within a particular model, but does not mean that the body will literally become younger by the specified number of years.

Let's imagine a 55-year-old person whose test result was 47.

The most cautious interpretation:

His set of measured biomarkers according to this algorithm appears more favorable than the average for reference 55-year-olds and is consistent with a profile typical of younger or lower-risk individuals.

It is not possible to conclude automatically:

• "he will live eight years longer";
• "all his organs are 47 years old";
• "his risk of all diseases is the same as that of the average 47-year-old";
• "In a year the test will definitely show 48."

Different watches can give different results to the same person precisely because they measure different biological signals. [31]

What does it mean to have biological age older than your passport age?

A result higher than the calendar age is not a diagnosis of a disease and should not lead to an attempt to “treat aging” on your own.

If a value is obtained by a validated model, it may indicate a less favorable profile of the biomarkers used by that algorithm. For example, some watches may be sensitive to inflammation, metabolic status, smoking, or other factors.

But the correct next question is:

"What specific health factors require attention?"

And not:

"How to quickly remove seven years from your biological age?"

If an elevated Phenotypic Age is associated with high glucose, C-reactive protein, or other abnormal test results, it is medically useful to evaluate these specific indicators according to standard clinical guidelines. The Phenotypic Age Index itself does not replace the diagnosis of diabetes, cardiovascular disease, inflammatory disease, or renal impairment.

Is it possible to measure biological age once and know the rate of aging?

No. A single measurement shows the state at a specific moment, but does not reliably determine an individual's trajectory.

Age and the rate of aging are different concepts.

If a test shows "47 years old" today, and a person is 50, we don't know from this number whether they are aging more slowly or faster than usual. Understanding the changes requires either an algorithm specifically designed to assess aging, such as DunedinPACE, or repeated comparable measurements. [32]

When repeating tests, it's advisable to use the same laboratory, the same sample type, and the same algorithm. Comparing today's GrimAge with last year's online blood test calculator is practically pointless: the change in method alone can significantly outweigh the biological effect of interest.

Even with the same test, a small change must be considered in light of technical and biological variability. This is one of the key unresolved issues in the individual use of clocks. [33]

Is it possible to "reverse biological age"?

Some biomarkers of aging do change in response to interventions, but a reduction in clock speed does not necessarily prove overall rejuvenation or life extension.

A good example is the randomized CALERIE trial. It followed 220 non-obese adults for two years after assigning either calorie restriction or a normal diet. Energy restriction slowed DunedinPACE by approximately 2-3% but did not significantly change PhenoAge or GrimAge. [34]

This result is both interesting and instructive.

It shows that the biomarker can respond to real intervention.

But it also shows that different clocks may not agree with each other.

And most importantly, the authors themselves emphasized that definitive proof of the geroprotective effect requires long-term studies with real clinical outcomes, including morbidity and mortality. [35]

A similar example was provided by DO-HEALTH. In an analysis of a randomized trial in older adults, omega-3, vitamin D, and exercise altered some epigenetic clocks, with the magnitude and presence of the effect depending on the specific algorithm. The authors specifically noted that there is no gold standard for biological aging. [36]

Therefore the phrase:

"The program rejuvenated participants by three months."

Much stronger than the data suggests.

More correctly:

"a specific epigenetic indicator changed in a direction that the model interprets as slower or younger aging."

Does lifestyle affect biological age?

Lifestyle is associated with a number of biomarkers of aging, but the evidence is not consistent across watches and does not allow each association to be translated into a number of "years lost."

A recent systematic review and meta-analysis from 2026 included 44 studies and over 145,000 participants. Higher levels of physical activity were associated with less age acceleration according to Horvath and GrimAge, but no statistically significant association was found for Hannum and PhenoAge in the meta-analysis. A major limitation was that most of the evidence came from cross-sectional observational studies, so causality remains uncertain. [37]

This nicely illustrates the general problem with clocks: physical activity is reliably beneficial for cardiovascular health, metabolism, muscle, and functional aging, regardless of whether a particular epigenetic test shows minus one year or no change.

That's why you shouldn't base a healthy lifestyle on the goal of "improving GrimAge." It's more rational to engage in physical activity, quit smoking, and control your blood pressure, lipids, and glucose metabolism, as these actions have direct, proven medical benefits.

Biological age may one day become an additional way to track effect, but should not replace these goals.

Is there a test that shows the most accurate biological age?

No. In 2026, such a test does not exist.

A systematic review of methods concluded that there is no agreed gold standard.[38]

The Biomarkers of Aging Consortium also emphasizes that further clinical translation requires systematic validation—the ability of a biomarker to reliably predict important outcomes, be reproducible in individuals, and respond to interventions in a way that truly informs medical decisions. [39]

Even within one popular class—the epigenetic clocks—a large 2026 study showed different molecular underpinnings for five models.[40]

Therefore, the question “which analysis is the most accurate?” is best reformulated for now:

Accurate for what task?

To guess the chronological age?

Predict mortality?

Assess the rate of aging?

Respond to intervention?

Describe the aging of a specific organ?

Different methods were optimized for different tasks.

Is it worth taking a commercial biological age test?

For a healthy person, such a test can be considered a research or wellness tool, but today it rarely changes standard medical practice.

This is especially important because the cost of a complex test can create a perception of high medical precision. Epigenetic analysis is indeed technologically complex and scientifically interesting, but the technology used to measure DNA and the clinical utility of the resulting results are two different issues.

In a 2025 review specifically focusing on the transition of epigenetic clocks from population science to the individual clinic, the authors concluded that current clocks do not yet meet the usual standards of clinical utility at the individual level and should not be used for independent decision making. [41]

This doesn't mean that commercial tests are useless or inherently fraudulent. Some use scientifically validated algorithms and can provide interesting information.

But the usefulness of the result is much higher if the laboratory transparently reports:

• what specific algorithm is used;
• what tissue is being analyzed;
• on what population the model was developed;
• what does the final number mean;
• What is the technical reproducibility;
• Is it possible to compare repeated measurements correctly?
• what clinical outcomes were predicted during validation;
• What is the uncertainty of the individual outcome?

If the user only gets a nice number "biological age 38.7 years" without these explanations, the medical interpretation is severely limited.

Is it possible to calculate biological age online?

It is possible to obtain a rough estimate, but most simple online calculators cannot be considered a complete measurement of biological aging.

The calculator can take your calendar age, body weight, smoking, sleep, physical activity level, and lab values, assign a specific number of "years" to each factor, and return a result.

The main problem isn't the idea of a mathematical model itself—scientific clocks are algorithms, too. The problem is how exactly this algorithm is created and validated.

If the coefficients are invented by the site creator or are based on disparate associations, the result does not become a biological age just because it is expressed in years.

Even among scientific models based on common clinical outcomes, a 2026 systematic review found a high incidence of methodological problems and insufficient external validation.[42]

Therefore, it is more useful to think of a simple calculator as an illustration of a lifestyle rather than as a laboratory test.

Is it possible to determine biological age by a face, a photograph or a watch?

Machine learning algorithms are indeed capable of assessing age patterns from images, electrocardiograms, magnetic resonance imaging, and other digital data. A modern review from 2026 identifies functional, molecular, and digital methods as actively developing areas of biological age measurement. [43]

However, the same rule applies here: the ability of an algorithm to predict age or medical outcome does not mean the existence of a universal digital “biological age.”

For example, a neural network might detect age-related changes in the heart on an electrocardiogram, while another might detect age-related features of the brain on an MRI. These models may be clinically useful for their respective purposes, but their results do not necessarily correlate.

Therefore, a person may appear younger than their official age, their cardiovascular system may have a favorable profile, but their metabolic indicators, conversely, may be more unfavorable. This is one of the fundamental reasons why the idea of a singular aging remains controversial.

What to do if the test shows high biological age

Don't try to "fix" a number. It's important to identify the specific measurable factors that contributed to the unfavorable outcome.

If a blood test model was used, it is worth looking at the laboratory results themselves.

If it's GrimAge or another epigenetic test, it's helpful to know what clinical risks the model actually validates and how reliable the individual result is.

After this, it is more reasonable to focus on traditional factors that have clear medical effects: blood pressure, lipids, glycemia, smoking, body weight and waist circumference in the appropriate context, physical activity, diseases, organ function and age-related preventive examinations.

This approach addresses the main unsolved problem that the Biomarkers of Aging Consortium identifies for the entire field: a biomarker must be linked to clinically useful information, otherwise a pretty number by itself is of little help to the patient. [44]

For example, if a test says "+7 biological years," and a routine examination reveals persistent high blood pressure, the medically significant action becomes treating the hypertension—not buying a supplement that promises to reduce epigenetic age.

What to do if the test shows a very young age

You can be happy with the result, but you shouldn’t consider it a medical guarantee.

Low biological age does not cancel out:

• age screening;
• blood pressure control;
• cardiovascular risk assessment;
• recommended vaccinations;
• examination for symptoms;
• treatment of already diagnosed diseases.

A person chronologically 60 years old with an epigenetic age of 50 is not considered medically 50 years old. Age according to one's passport remains one of the strongest risk factors for many diseases and continues to be used in clinical guidelines.

The biological clock can add information to calendar age, but does not yet replace it. [45]

How to compare repeated results correctly

If a person nevertheless decides to track biological age over time, it is most reasonable to follow several principles.

First of all, you need to use the same method. The change from GrimAge to PhenoAge cannot be interpreted as a change.

It is advisable to use the same tissue and laboratory platform, as technical differences may influence the results of epigenetic models.[46]

It is not advisable to draw conclusions from a small change without knowing the analytical variability of the test.

And most importantly, the dynamics of clocks must be compared with changes in real health: blood pressure, physical fitness, laboratory risk factors, functional status, and diseases.

If all the standard indicators improve, and one experimental model becomes “a year older,” this is not a reason to automatically change a healthy lifestyle.

What is often misunderstood

"Biological age is the actual age of the body."

Not exactly. It is a calculated indicator, the definition of which depends on a set of biomarkers and a model. There is no single gold standard. [47]

"If I'm 50 and the test shows 40, I'll live 10 years longer."

Such a conclusion cannot be drawn. Some models are associated with mortality risk, but the difference in years they predict does not directly translate into additional years of life.

"The epigenetic clock accurately reveals true biological age"

They are one of the most studied tools in geroscience, but even epigenetic clocks differ significantly from each other and reflect different molecular processes. [48]

Horvath Age, GrimAge, and PhenoAge are the same thing.

No. The first Horvath watches were primarily trained to predict chronological age, while PhenoAge and GrimAge were designed with a greater focus on health, risk, and survival.

"Telomeres are the most accurate age counter"

No. Telomere length is only moderately associated with age and varies greatly depending on the method and tissue. In the largest meta-analysis, the correlation with chronological age was relatively weak. [49]

"If biological age improves after a diet, rejuvenation is proven."

No. In CALERIE, caloric restriction affected DunedinPACE but did not change PhenoAge or GrimAge; the change in the surrogate biomarker itself remains to be confirmed by real-world long-term clinical outcomes. [50]

"A commercial epigenetic test could be used to guide treatment."

Current evidence for such use is insufficient. A separate critical review from 2025 considers individual decisions based on modern epigenetic clocks premature. [51]

What is the most useful assessment method today?

For the average person, practical value depends on the purpose.

Target	Which is more informative?
Understand your overall health	standard preventive examination and risk factors
Assess functional aging	strength, mobility, physical performance, cognitive state
Scientifically assess molecular aging	validated epigenetic clock
To assess the current rate of aging in a research context	DunedinPACE and similar pace models
Get an inexpensive research index	validated clinical biochemical clock
To study the aging of individual organs	developing proteomic/imaging models
Make a decision about treatment	standard clinical indicators and guidelines, not age-specific hours

This isn't a ranking of methods based on "accuracy." Each solves its own problem.

Key points from experts

Daniel W. Belsky, PhD - Associate Professor of Epidemiology, Robert N. Butler Columbia Aging Center, Columbia University Mailman School of Public Health; Director of the Geroscience Computational Core. His group develops methods for measuring the rate and progression of biological aging in humans and their application to epidemiology and clinical research. [52]

Belsky's research group developed DunedinPACE, which differs fundamentally from many age-based timepieces in that the model was trained on long-term changes in 19 physiological parameters in the same individuals. This highlights the important distinction between the body's cumulative state and its current rate of change. [53]

Tony Wyss-Coray, PhD - DH Chen, Distinguished Professor of Neurology and Neurological Sciences and Director of the Phil and Penny Knight Initiative for Brain Resilience, Stanford University. His lab researches brain aging, neurodegeneration, and systemic factors that influence age-related decline. [54]

A Nature Medicine 2026 review by Wyss-Coray and Eric Topol explores biological clocks as a promising infrastructure for assessing the aging of individual organs, tissues, and cells, predicting disease, and researching interventions. However, the field is still in the transition from strong research associations to clinically applicable individual metrics. [55]

Eric J. Topol, MD, is a cardiologist, Executive Vice President of Scripps Research, founder and director of the Scripps Research Translational Institute, and the Gary and Mary West Chair of Innovative Medicine. His work focuses on genomics, digital medicine, and personalized health assessment methods. [56]

A joint review by Topol and Wyss-Coray shows how rapidly the field of biological clocks has expanded from DNA methylation to proteomics, imaging, digital and organ-specific models. The practical implication of this evolution is not the emergence of a single definitive "aging test," but rather the recognition of the multidimensionality of the aging process. [57]

Frequently Asked Questions

Is it possible to find out your biological age from one blood test?

It is possible to calculate a specific index, but not the only true age of the organism. There are validated models, such as Phenotypic Age, that combine several blood indicators with calendar age. [58]

Which analysis is the most scientific?

Epigenetic clocks are among the most studied molecular approaches, but even among them there is no single best test for all purposes. [59]

Is it possible to determine biological age from DNA?

Yes, epigenetic clocks use DNA methylation patterns. But they analyze epigenetic marks, not "gene age," and the results depend on the specific algorithm. [60]

Which is better - Horvath or GrimAge?

They address different objectives. Horvath was created primarily as a tissue age predictor, while GrimAge was created as an indicator more closely related to survival and certain diseases. [61]

Which is better - GrimAge or DunedinPACE?

One cannot be universally considered the best. GrimAge focuses more on accumulated risk, while DunedinPACE focuses on the rate of physiological aging, reduced to an epigenetic indicator. [62]

Is it possible to be biologically older than your age?

Yes, within the specific model. This means a less favorable profile of the relevant biomarkers compared to peers, rather than a literal physical aging of the specified number of years.

Is it possible to be biologically younger?

Yes. But a difference of, say, -8 years does not automatically mean eight additional years of life.

How often does it make sense to measure biological age?

There is no universal medical interval. For research or personal monitoring, it is more important to use the same validated test and consider its reproducibility than to frequently repeat measurements.

Do telomere tests work?

They measure a real biological property associated with aging, but telomere length is too variable to be considered an accurate universal measure of an organism's age.[63]

Is it possible to calculate age using a smartwatch?

It is possible to create an algorithm based on heart rate, activity, sleep, and other digital indicators, but such a result remains a model, not a direct medical measurement of biological age.

Is it possible to reduce biological age with exercise?

Physical activity is associated with more favorable readings of some epigenetic clocks. A 2026 meta-analysis found such links for Horvath and GrimAge, but the evidence was primarily observational and does not allow one to definitively state the extent to which exercise "rejuvenates the clock" causally. [64]

Is it possible to reduce it with nutrition?

Some interventions alter individual aging clocks, but the results depend on the model. CALERIE, for example, affected DunedinPACE, but not PhenoAge and GrimAge. [65]

Can stress make a person biologically older?

Some biomarkers of aging are stress-related and can change temporarily. This is another reason not to interpret a single number as a constant "true age" of the body. The Biomarkers of Aging Consortium includes sensitivity to temporary stressors among the issues important for the clinical interpretation of such tests. [66]

Should I be concerned if my biological age is five years higher than my passport age?

It's not just the number itself. First, you need to understand the method and evaluate specific health factors. The aging clock result itself is not a diagnosis. [67]

Is it possible to prescribe dietary supplements based on test results?

There is no compelling reason to select supplements solely for the purpose of reducing biological age. Clinical decisions should be based on specific indications and validated medical indicators, not on the aging clock alone. [68]

Main

Biological age is a useful scientific concept, but today it is not a single measurable characteristic of the body. There are numerous clocks using blood, DNA methylation, proteins, functional indicators, and digital data. They can predict important age-related outcomes better than calendar age alone, but there is no universal gold standard yet. [69]

The most serious mistake is to interpret the result as a literal number of years the body has lived. A number like "42 years" is the output of a specific algorithm. Horvath, GrimAge, PhenoAge, and DunedinPACE may diverge because they reflect different biological processes. [70]

Therefore, for the average person, the aging clock is best viewed as supplemental research information rather than a replacement for standard preventative medicine. Blood pressure, glucose, lipids, medical conditions, physical function, smoking, diet, and activity are much easier to link to specific actions. The clinical challenge for the future is to prove when the biological age result adds sufficiently useful information to these data to truly change the prognosis or medical management. [71]