Biological age tests: methods, accuracy, and limits

What is a biological age test?

A biological age test uses molecular or clinical markers to estimate how old your cells and tissues appear according to a particular model. Most tests analyze blood or saliva and return either an age estimate or an estimated pace of aging. These are biomarkers of aging, not diagnostic tests: the number describes a pattern in the measurements, rather than establishing how healthy you are. Source

Aging is not one process. Epigenetic changes, shortening telomeres, chronic inflammation, and cellular senescence interact with one another. A widely cited 2023 framework describes 12 hallmarks of aging. No single consumer test directly measures all of them. Different biological targets, training methods, and sample types help explain why tests disagree. The framework itself is a way to organize research, not a universal checklist that a laboratory can turn into one definitive age. Source

You may see these products described as epigenetic age tests, aging clocks, longevity panels, or pace-of-aging tests. The names cover different methods with different supporting evidence. There is no single standard biological age test.

Biological age versus chronological age

Chronological age is the time since your birth. Biological age aims to describe aspects of aging that may vary between people born on the same day. Their test results can therefore differ by years. Neither a routine clinical measurement nor an aging clock captures every underlying aging mechanism, and a clock’s estimate is not a direct measurement of the total aging accumulated in your body. Source

Researchers often examine the gap between a model’s estimated age and chronological age. In one study, a methylation age five years higher than expected was associated with a 21% higher mortality risk after adjustment for age and sex. The association was 16% after further adjustment for IQ, education, social class, hypertension, diabetes, cardiovascular disease, and APOE status. These are comparisons between groups, not predictions of an individual’s lifespan. Source

How biological age tests work

A test measures a defined set of markers in a sample. An algorithm developed using data from groups of people converts the measurements into an age in years or a pace of aging. The three main approaches use DNA methylation patterns, telomere length, or routine blood chemistry.

The process usually involves these steps:

• Choose a method and provider, including the particular clock the laboratory uses.

• Follow the provider’s instructions for preparing your sample.

• Supply a venous blood sample, dried blood spot, saliva sample, or cheek swab, depending on the method.

• The laboratory measures the markers and applies its algorithm.

• Discuss the result alongside your other health information with a clinician who understands what the test can and cannot show.

Two tests can produce different results for the same person without either result representing a change in their health. Clocks trained to predict chronological age and clocks trained to predict other aspects of biological aging use different CpG patterns and answer different questions. Source

Method	What it measures	Sample type	Main use	Important limitation
DNA methylation	Methylation marks at CpG sites	Blood, saliva, or cheek swab, depending on validation	Research age estimates; some clocks have substantial validation	Validation does not automatically extend to every commercial product, tissue, or clock version
Telomere length	Length of chromosome end caps	Usually white blood cells	Population research	Weak correlation with chronological age
Blood biomarkers, or phenotypic age	Routine laboratory markers from several organ systems	Blood draw	Repeatable estimates using standard laboratory measurements	Reflects physiological measurements rather than DNA methylation
Questionnaire estimates	Self-reported habits and medical history	Questionnaire	A quick way to consider health habits; some tools are free	Does not directly measure biological aging

DNA methylation and epigenetic clocks

An epigenetic clock reads chemical methylation marks at selected CpG sites on DNA, then converts their pattern into an estimate. The original multitissue clock used about 8,000 samples from 82 datasets covering 51 healthy tissues and cell types. Its calculation used 353 CpG sites. Modern commercial tests draw on the broader clock concept, but their algorithms, laboratory methods, and validation can differ considerably from that original model. Source

Telomere length

Telomeres are protective caps at the ends of chromosomes. A telomere test measures their length, usually in white blood cells. A single measurement does not reveal how much that person’s telomeres have shortened since an earlier point in life. The association with age is modest: across 743,019 people, the pooled correlation between telomere length and chronological age was −0.19 and weakened at older ages. In a separate study of 474,074 adults, age, ethnicity, sex, and white blood cell count together explained only about 5.5% of the variation in telomere length. The two papers address different questions: one summarizes the relationship with age, while the other characterizes variation in UK Biobank. Source Source

Blood markers and phenotypic age

Clinical Phenotypic Age combines chronological age with nine routine blood biomarkers; it does not require DNA sequencing or a methylation assay. Its US validation included 11,432 adults ages 20–84, plus 185 participants whose ages were recorded as 85 or older. Researchers observed 1,012 deaths over 12.6 years of follow-up. This approach uses familiar clinical measurements to produce an age-related risk estimate. Access to routine tests may make repeat measurement more practical, but that does not establish the usefulness of repeatedly buying a consumer age score. Source

What the main aging clocks measure

Aging clocks are not interchangeable. Early clocks were trained primarily to predict chronological age. Later models were trained on clinical measures associated with aging or on mortality-related outcomes. Even similar names can describe different tests: clinical Phenotypic Age uses blood chemistry, while DNAm PhenoAge is a methylation-based predictor trained to estimate that clinical measure. Treating all these outputs as one universal “biological age” hides those differences. Source

Clock	What it estimates	Typical research use
Horvath	Chronological age across multiple tissues	Comparing methylation age with calendar age
DNAm PhenoAge	A methylation-based estimate trained on clinical Phenotypic Age	Studying associations with mortality, healthspan, and physical function
GrimAge	A mortality-related composite score derived from methylation proxies	Studying population associations with survival, heart disease, and cancer; not predicting an individual’s death date
DunedinPACE	A pace-of-aging measure	Studying variation in aging pace and responses to interventions

DunedinPACE reports a pace rather than an age. Its training target was based on changes in 19 organ-system indicators measured in members of one birth cohort at four time points over two decades. A speedometer is a useful analogy: it aims to estimate the rate of change rather than the total distance traveled. GrimAge uses a different approach, combining methylation-derived proxies for seven plasma proteins and smoking exposure measured in pack-years. Both models are useful research tools, but their outputs are not equivalent. Source Source

Which biological processes does a test capture?

The answer depends on what was measured: methylation marks on DNA, chromosome end-cap length, or routine blood markers. Each captures part of aging biology, often indirectly. None is a direct, complete measurement of all the processes involved in aging. That does not make the information meaningless; it means the interpretation needs to stay within the method’s scope. Source

DNA methylation is widely used in aging-clock research. Over time, methylation marks are gained or lost at particular positions on DNA. Some of these patterns are sufficiently predictable to support age estimation. Think of the clock as reading selected addresses on DNA and applying a calculation. A model validated across many tissues may work differently from one developed for blood alone, so the original multitissue findings do not validate every clock in every sample type. Source

Blood-based clinical scores take another route. Measurements related to liver function, blood sugar regulation, kidney function, inflammation, and other physiological features contribute to an age-related estimate. Their relationships with mortality can be tested across populations. They provide a different view of health from a methylation assay, rather than a cheaper measurement of precisely the same thing. Source

How accurate are biological age tests?

Several clocks identify meaningful differences in outcomes between groups. That is a different achievement from precisely measuring one person’s aging or predicting their lifespan. Consumer biological age tests are not diagnostic tests and are not established as part of routine clinical care.

Repeatability is another concern. In a study of six widely used epigenetic clocks, technical noise produced differences of up to nine years between repeat measurements of the same sample. Versions based on principal component analysis reduced differences to about 1.5 years for most replicates. Better repeatability is valuable, but it is not the same as proving that a number is a person’s true biological age. The uncertainty depends on the clock and its version. Source

What the research supports

A systematic review and meta-analysis covering 23 studies and 41,607 participants found that each five-year increase in DNA methylation age was associated with an 8–15% higher mortality risk. This supports the value of these measurements for studying population risk. It does not establish that changing a person’s clock reading by five years will change their mortality risk by the corresponding amount. Source

A separate analysis across multiple cohorts also found that methylation-based age measures predicted all-cause mortality. Measures incorporating information about blood cell composition improved prediction. These findings show that the measurements contain useful information, while also illustrating that the construction of the score matters. Source

What remains uncertain

The same systematic review found less consistent evidence for specific age-related diseases and longevity. It also identified probable publication bias favoring positive findings. A strong association with overall mortality therefore should not be extended automatically to every disease or to a claim that the test can guide treatment. Source

A nationally representative study of older US adults adds perspective. Accelerated aging measured with PhenoAge, GrimAge, or DunedinPACE predicted cognitive impairment, functional limitations, chronic disease, and four-year mortality. However, demographic characteristics, socioeconomic status, mental health, and health behaviors were similarly strong or stronger predictors. An aging clock can add another piece of information, but it does not replace the person’s circumstances and clinical history. Source

How to take a biological age test

Start with the method, not the promise on the package. Find out which clock the provider uses, what sample it requires, and what evidence supports its use in people like you. Follow the collection instructions and interpret the report with your other health information. A number without that context can be easy to overread.

Before choosing a test, keep four points in mind:

• The result is not a diagnosis, and consumer aging clocks are not established routine clinical tests.

• Different clocks can assign different ages to the same person because their targets and CpG patterns differ. Source

• Repeat measurements of the same sample have differed by as much as nine years from technical noise alone in a study of several clocks. That is not a universal error margin for every product. Source

• Telomere length alone is a limited marker of biological aging. One length measurement should not be treated as a direct measurement of your current rate of aging. Source

Blood or saliva: at home or in a clinic?

Sample type can substantially affect the answer. A study of 284 samples from 83 people ages 9–70 compared cheek cells, saliva, dried blood spots, buffy coat, and peripheral blood mononuclear cells. For some clocks, average within-person differences between oral and blood estimates approached 30 years. This was a discrepancy between model estimates in different tissues, not evidence that one part of the person’s body was literally 30 years older. Source

Another study applied blood-derived algorithms to saliva. Cross-tissue agreement was moderate for some second- and third-generation measurements and poorer for first-generation age-acceleration measures. These results support choosing a sample type validated for the specific clock. They do not show that every saliva test is invalid or every blood test is better. For repeat measurements, keep the sample type and provider consistent. Source

Preparing for the test

Consistent preparation can reduce avoidable differences, but much of the technical variation is outside your control. Following instructions will not eliminate laboratory noise. The finding that identical samples can produce different readings is a reason to interpret small changes cautiously, not a reason to blame preparation for every discrepancy. Before collection: Source

• Follow fasting instructions if the particular test requires fasting.

• Use a similar time of day when this is relevant to the markers being measured, especially for routine blood panels.

• If you have an acute illness, ask whether collection should be postponed rather than assuming the result represents your usual state.

• Keep the collection method and provider the same across repeat tests.

What does a biological age test cost?

Ask for the current total price, including collection, laboratory analysis, interpretation, and any repeat testing. Prices and coverage depend on the provider, method, and location; the papers validating clocks are not price comparisons. Clinical Phenotypic Age uses routine blood measurements rather than a methylation assay, which may allow use of results already available. That methodological distinction does not establish a universal price ranking or insurance policy. Source

Interpreting your result

An estimated age is usually compared with chronological age. A pace-of-aging test reports a rate instead. The interpretation depends on how the model was trained and validated. A population association is not a personal prognosis, and a lower number is not automatically a better treatment target.

Result pattern	What it may suggest	What it does not establish
Estimated age below chronological age	A comparatively favorable pattern within the particular model	Proof that you will live longer
Approximately equal	A result near the model’s age-related reference	Confirmation of overall good health
Estimated age above chronological age	A pattern associated with higher risk in some validated models	A diagnosis or prediction of your future
Pace above 1.0	Faster change relative to the model’s reference pace	That every organ is literally aging at that rate, or that the rate is permanent

The pace interpretation comes from the model’s reference and development, not from directly observing every aging process in your body. Source

Read the result alongside clinical measurements and medical history. Health behaviors, socioeconomic circumstances, and other familiar information can be as predictive of aging-related outcomes as a clock, or more so. Context is what turns an isolated number into information worth discussing. Source

What can change a biological age measurement?

Some interventions have changed aging-clock measurements in randomized trials. The effects are generally modest and vary by clock. These findings are promising for research, but a change in a biomarker does not by itself demonstrate a longer life or better health. It also does not tell us which component of a combined intervention produced the change.

• Calorie restriction: CALERIE randomized 220 adults without obesity to a target of 25% calorie restriction or unrestricted eating for two years. The methylation analysis included 197 participants with relevant samples. Restriction modestly slowed DunedinPACE but did not significantly change the PhenoAge or GrimAge measures used. The 25% figure was a prescribed target, not the average reduction achieved. This trial is not a recommendation to cut your own intake by that amount. Source

• Diet and physical activity: The 24-month DAMA factorial trial included 219 healthy postmenopausal women. The dietary intervention slowed GrimAge, while the physical activity intervention reduced stochastic epigenetic mutations. These were different outcomes; the findings do not mean that both interventions reduced GrimAge. Source

• Polyphenol intake: In the 18-month DIRECT PLUS study of 256 adults with abdominal obesity or dyslipidemia, greater adherence to the green Mediterranean diet and higher polyphenol exposure were associated with some more favorable clock changes. However, the randomized dietary groups did not differ significantly in clock changes. The adherence associations should not be presented as proof that polyphenols caused slower aging. Source

• A combined lifestyle program: An eight-week pilot trial randomized 43 men ages 50–72 to a program combining diet, sleep, exercise, relaxation advice, and supplements or to control. The between-group difference in Horvath methylation age was 3.23 years. The intervention group’s own change from baseline was not statistically significant. This small, preliminary trial does not demonstrate reversal of aging or identify which part of the program mattered. Source

Observational studies also link physical activity with more favorable clock readings. In the Rhineland study, accelerometer-measured activity among 3,567 adults ages 30–94 was associated with slower epigenetic aging across several clocks. Among 2,435 Framingham participants, 1,500 additional daily steps or three fewer sedentary hours were associated with more than 10 months lower GrimAge. After adjustment for blood cell composition, the difference fell to about one month. These are associations, not proof that adding those steps will lower your age by a fixed amount. For a practical activity target, US guidelines recommend 150–300 minutes of moderate aerobic activity a week plus muscle strengthening on two days. A clock study cannot set an individualized step prescription. Source Source Source

How often should you repeat a test?

There is no validated universal retesting interval for consumer biological age tests. Testing every few weeks may produce changes that are difficult to distinguish from measurement noise. Waiting months or a year is sometimes proposed as a practical approach, but it does not guarantee that a change exceeds the test’s uncertainty. Repeat measurements of the same sample have differed by several years with some clocks. First consider what decision a repeat result would actually inform. Source

Keep the provider, clock version, sample type, and collection conditions consistent where possible. Cross-tissue results cannot simply be substituted for one another. Intervention studies have used periods ranging from eight weeks to two years, but those research schedules do not establish a clinical retesting recommendation. Source

Who might find a biological age test useful?

A test may interest someone who already monitors their health and wants to follow an additional research-derived measure. It is reasonable to approach it as an optional source of information, provided its limits and cost are acceptable. If blood pressure, routine health needs, sleep, and physical activity have not yet received attention, an aging score is unlikely to be the most useful first step. Demographic and socioeconomic factors, mental health, and health behaviors can predict aging-related outcomes as well as or better than clock measures. Source

These tests should not be used to explain persistent symptoms or to decide whether symptoms deserve medical attention. They are not diagnostic tests. A favorable longevity-panel result does not replace assessment of a health concern.

Frequently asked questions

How accurate are biological age tests?

Their strongest evidence concerns differences between groups, rather than precise individual ages. Across 23 studies with 41,607 participants, each five-year increase in methylation age was associated with an 8–15% higher mortality risk. Repeatability is a separate issue: technical noise produced differences of up to nine years between identical-sample measurements with six clocks. Principal-component versions reduced the difference to roughly 1.5 years for most replicates. Neither figure describes the accuracy of every commercial test, and better repeatability is not proof of a true biological age. Source Source

What is the difference between biological and chronological age?

Chronological age is time since birth. Biological age is an estimate based on selected measurements and a particular model, not a direct count of aging accumulated in cells. In one study, a five-year higher methylation age was associated with 21% higher mortality risk after adjustment for age and sex, and 16% after further adjustment for factors including education, social class, and cardiometabolic conditions. Those population findings do not predict an individual’s remaining years. Source Source

How does a biological age test work?

It measures selected markers in a sample and applies an algorithm developed using data from groups of people. The main approaches use DNA methylation, telomere length, or routine blood chemistry. The original multitissue epigenetic clock used 353 CpG sites across 51 healthy tissue and cell types. Clinical Phenotypic Age instead combines nine routine blood biomarkers with chronological age. The output therefore depends on both what is measured and what the algorithm was designed to estimate. Source Source

Why do two tests give me different ages?

Different clocks have different training targets and CpG patterns. Sample type adds another source of variation. In a study of 284 samples from 83 people, average differences between oral and blood estimates approached 30 years for some clocks. A separate study found only moderate agreement for some second- and third-generation blood-derived measures applied to saliva. Compare the same method and sample type over time, and do not interpret disagreement between tests as proof of a rapid health change. Source Source Source

Is a blood or saliva test better?

The useful question is whether the particular clock has been validated for the sample being collected. A blood-trained algorithm is not automatically valid in saliva. Studies have found substantial within-person differences between oral and blood estimates and poor cross-tissue agreement for some first-generation age-acceleration measures. Those findings do not establish a universal ranking of every blood and saliva product. Ask about validation and keep the provider, method, and sample type consistent when repeating a test. Source Source

Can you lower your measured biological age?

Some trials have shifted clock measurements, but different clocks respond differently. CALERIE randomized 220 adults to a 25% calorie-restriction target or unrestricted eating; the methylation analysis in 197 participants found a small DunedinPACE effect, without significant PhenoAge or GrimAge change. DAMA found a dietary effect on GrimAge in 219 postmenopausal women. DIRECT PLUS found associations with dietary adherence and polyphenol exposure among 256 adults, but no significant difference in clock changes between randomized diet groups. These findings do not yet show that lowering a consumer clock score extends life. Source Source Source

Is a telomere test a good measure of biological age?

Telomere length is a limited standalone marker. Its pooled correlation with chronological age was only −0.19 across 743,019 people and weakened at older ages. In 474,074 UK Biobank participants, age, ethnicity, sex, and white blood cell count explained about 5.5% of the variation in telomere length. The biology remains relevant to aging research, but a single length measurement is not a reliable reading of how quickly you are aging now. Source Source Source

How often should you retest your biological age?

There is no established universal schedule. Months or a year may be more practical than frequent short-interval testing, but a longer interval does not ensure that a measured change is real. Technical differences between repeat measurements can be substantial. Research interventions lasting eight weeks to two years do not establish a consumer retesting guideline. If repeating a test has a clear purpose, keep the provider, clock version, sample type, and collection conditions consistent and interpret the result with its uncertainty. Source Source

Sources and references

Horvath S, Raj K (2018). DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nature Reviews Genetics, 19(6), 371–384. https://doi.org/10.1038/s41576-018-0004-3 Publication

López-Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G (2023). Hallmarks of aging: an expanding universe. Cell, 186(2), 243–278. https://doi.org/10.1016/j.cell.2022.11.001 Publication

Marioni RE, Shah S, McRae AF, et al (2015). DNA methylation age of blood predicts all-cause mortality in later life. Genome Biology, 16(1), 25. https://doi.org/10.1186/s13059-015-0584-6 Publication

Bernabeu E, McCartney DL, Gadd DA, et al (2023). Refining epigenetic prediction of chronological and biological age. Genome Medicine, 15(1), 12. https://doi.org/10.1186/s13073-023-01161-y Publication

Horvath S (2013). DNA methylation age of human tissues and cell types. Genome Biology, 14(10), R115. https://doi.org/10.1186/gb-2013-14-10-r115 Publication

Ye Q, Apsley AT, Etzel L, et al (2023). Telomere length and chronological age across the human lifespan: a systematic review and meta-analysis of 414 study samples including 743,019 individuals. Ageing Research Reviews, 90, 102031. https://doi.org/10.1016/j.arr.2023.102031 Publication

Codd V, Denniff M, Swinfield C, et al (2022). Measurement and initial characterization of leukocyte telomere length in 474,074 participants in UK Biobank. Nature Aging, 2(2), 170–179. https://doi.org/10.1038/s43587-021-00166-9 Publication

Liu Z, Kuo PL, Horvath S, Crimmins E, Ferrucci L, Levine M (2018). A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV: a cohort study. PLoS Medicine, 15(12), e1002718. https://doi.org/10.1371/journal.pmed.1002718 Publication

Levine ME, Lu AT, Quach A, et al (2018). An epigenetic biomarker of aging for lifespan and healthspan. Aging, 10(4), 573–591. https://doi.org/10.18632/aging.101414 Publication

Belsky DW, Caspi A, Corcoran DL, et al (2022). DunedinPACE, a DNA methylation biomarker of the pace of aging. eLife, 11, e73420. https://doi.org/10.7554/eLife.73420 Publication

Lu AT, Quach A, Wilson JG, et al (2019). DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging, 11(2), 303–327. https://doi.org/10.18632/aging.101684 Publication

Higgins-Chen AT, Thrush KL, Wang Y, et al (2022). A computational solution for bolstering reliability of epigenetic clocks: implications for clinical trials and longitudinal tracking. Nature Aging, 2(7), 644–661. https://doi.org/10.1038/s43587-022-00248-2 Publication

Fransquet PD, Wrigglesworth J, Woods RL, Ernst ME, Ryan J (2019). The epigenetic clock as a predictor of disease and mortality risk: a systematic review and meta-analysis. Clinical Epigenetics, 11(1), 62. https://doi.org/10.1186/s13148-019-0656-7 Publication

Chen BH, Marioni RE, Colicino E, et al (2016). DNA methylation-based measures of biological age: meta-analysis predicting time to death. Aging, 8(9), 1844–1865. https://doi.org/10.18632/aging.101020 Publication

Faul JD, Kim JK, Levine ME, Thyagarajan B, Weir DR, Crimmins EM (2023). Epigenetic-based age acceleration in a representative sample of older Americans: associations with aging-related morbidity and mortality. Proceedings of the National Academy of Sciences, 120(9), e2215840120. https://doi.org/10.1073/pnas.2215840120 Publication

Vaiserman A, Krasnienkov D (2021). Telomere length as a marker of biological age: state-of-the-art, open issues, and future perspectives. Frontiers in Genetics, 11, 630186. https://doi.org/10.3389/fgene.2020.630186 Publication

Apsley AT, Ye Q, Caspi A, et al (2025). Cross-tissue comparison of epigenetic aging clocks in humans. Aging Cell, 24(4), e14451. https://doi.org/10.1111/acel.14451 Publication

Zarandooz S, Raffington L (2025). Applying blood-derived epigenetic algorithms to saliva: cross-tissue similarity of DNA-methylation indices of aging, physiology, and cognition. Clinical Epigenetics, 17(1), 61. https://doi.org/10.1186/s13148-025-01868-2 Publication

Waziry R, Ryan CP, Corcoran DL, et al (2023). Effect of long-term caloric restriction on DNA methylation measures of biological aging in healthy adults from the CALERIE trial. Nature Aging, 3(3), 248–257. https://doi.org/10.1038/s43587-022-00357-y Publication

Fiorito G, Caini S, Palli D, et al (2021). DNA methylation-based biomarkers of aging were slowed down in a two-year diet and physical activity intervention trial: the DAMA study. Aging Cell, 20(10), e13439. https://doi.org/10.1111/acel.13439 Publication

Yaskolka Meir A, Keller M, Hoffmann A, et al (2023). The effect of polyphenols on DNA methylation-assessed biological age attenuation: the DIRECT PLUS randomized controlled trial. BMC Medicine, 21(1), 364. https://doi.org/10.1186/s12916-023-03067-3 Publication

Fitzgerald KN, Hodges R, Hanes D, et al (2021). Potential reversal of epigenetic age using a diet and lifestyle intervention: a pilot randomized clinical trial. Aging, 13(7), 9419–9432. https://doi.org/10.18632/aging.202913 Publication

Fox FAU, Liu D, Breteler MMB, Aziz NA (2023). Physical activity is associated with slower epigenetic ageing — findings from the Rhineland study. Aging Cell, 22(6), e13828. https://doi.org/10.1111/acel.13828 Publication

Spartano NL, Wang R, Yang Q, et al (2023). Association of accelerometer-measured physical activity and sedentary time with epigenetic markers of aging. Medicine & Science in Sports & Exercise, 55(2), 264–272. https://doi.org/10.1249/MSS.0000000000003041 Publication

US Department of Health and Human Services (2018). Physical activity guidelines for Americans (2nd ed.). US Department of Health and Human Services. https://odphp.health.gov/sites/default/files/2019-09/Physical_Activity_Guidelines_2nd_edition.pdf Publication