Healthy Ageing & Longevity - Longevity claims
Biological-age and epigenetic tests
Last updated 30 September 2026. How we write this .
A biological-age test estimates how old your body seems from chemical tags on your DNA. The best clocks predict health risk across large groups, but one person's result can shift by years between repeat tests.
Start here A biological-age test estimates how old your body seems from chemical tags on your DNA. The best clocks predict health risk across large groups, but one person's result can shift by years between repeat tests.
By the end, you will understand What does an epigenetic clock actually measure? What are PhenoAge, GrimAge and DunedinPACE? If I take the test twice, will I get the same answer? Key words used in this lesson Sequence The specific order of amino acids in a peptide. The sequence determines the shape the molecule takes and therefore what it can bind to, which is why changing one amino acid can change a compound entirely. Full plain-English glossary
Where this information comes from Claim-cited evidence. Central research claims on this lesson are linked to specific studies or official guidance.
This lesson links the specific studies and guidance used for its central claims.
Our standard: expert or athlete opinion may illustrate a viewpoint, but it is never treated as proof. Read the editorial policy .
Report an error or outdated claim .
A biological-age test estimates how old your body seems from chemical tags on your DNA, then reports it as an age or a pace of ageing. The best of these measures predict illness and death well across large groups of people. For one person, a single result is less reliable than the marketing suggests, and no trial yet shows that lowering your score improves your health.
What does an epigenetic clock actually measure?
Your DNA sequence stays almost the same for life, but the chemical tags attached to it change. The best-studied tag is DNA methylation, a small methyl group fixed to the DNA at particular spots called CpG sites. Methylation helps switch genes on and off, and at some sites it drifts in a predictable way as we get older. An epigenetic clock is a formula that reads methylation at a few hundred of these sites and turns the pattern into a number.
The best-known early clock came from Steve Horvath of the University of California, Los Angeles, published in Genome Biology in 2013. He built it from 8,000 samples covering 51 healthy tissues and cell types, and it reads 353 CpG sites. In test data its estimate tracked real age with a correlation of 0.96, and for half of people it landed within 3.6 years of their actual age. It read close to zero in embryonic stem cells, and an average of 36 years too old in cancer tissue.
Clocks like Horvath's were trained to predict calendar age. That makes them good age estimators, but researchers wanted something that tracked health instead. So the next generation was trained on different targets.
What are PhenoAge, GrimAge and DunedinPACE?
PhenoAge, published in 2018 by Morgan Levine and colleagues, was trained on a score built from routine clinical measures linked to illness and death, rather than on calendar age. It beat earlier clocks at predicting death from any cause, cancers, physical function and Alzheimer's disease.
GrimAge, from Ake Lu, Horvath and colleagues in 2019, went further. It uses methylation patterns that stand in for seven blood proteins and for a person's lifetime smoking, and combines them into an estimate tied to time until death. The team tested it on blood from 6,935 people in several large US and Italian studies. It stood out among existing clocks for predicting time to death, heart disease and cancer.
DunedinPACE reports a speed rather than an age. Daniel Belsky and colleagues used the Dunedin Study, which has followed 1,037 babies born in one New Zealand city in 1972 and 1973. Researchers tracked 19 measures of heart, metabolic, kidney, liver, immune, dental and lung health at ages 26, 32, 38 and 45, and worked out how fast each person was declining. The average was one year of biological ageing per calendar year, but the slowest person aged 0.40 years per year and the fastest 2.44. DunedinPACE is a blood test designed to estimate that pace from a single sample, so a score of 1.1 suggests ageing about 10% faster than average. In other groups of people it predicted illness, disability and death about as well as GrimAge.
If I take the test twice, will I get the same answer?
Not always, and this is the weakest point for personal use. In 2022, Albert Higgins-Chen and colleagues reported in Nature Aging what happens when the same DNA is measured more than once. Technical noise alone, with no real change in the person, shifted results by up to 9 years for six well-known clocks. Their fix was a statistical rebuild of each clock, which brought most repeat measurements within 1.5 years of each other.
DunedinPACE was built to avoid unreliable sites from the start. In repeat measurements of the same blood samples, its agreement score, called the intraclass correlation, was 0.96 on a scale where 1 means perfect agreement. That fell to 0.87 when the two runs used different versions of the laboratory chip.
For anyone buying a test, the practical upshot is simple. A drop of two or three years between tests may be nothing but noise. Unless a company says which clock, which version and which laboratory method it used, you cannot judge how much of a change is real. Results from different clocks are not interchangeable either, because each was trained to answer a different question.
Can a lifestyle change or treatment make my biological age younger?
The strongest evidence so far comes from CALERIE, a randomised trial in which 220 healthy adults without obesity were assigned for two years either to aim for 25% fewer calories or to eat as normal. A later analysis of their blood, published in Nature Aging in 2023, found the calorie-restriction group had a slower pace of ageing on DunedinPACE. PhenoAge and GrimAge did not change significantly, and the effects overall were small. One intervention, measured with different clocks, gave different answers.
The bigger gap is that nobody has yet shown that pushing a clock down leads to fewer heart attacks, less dementia or a longer life. Steven Cummings and Stephen Kritchevsky made this point in GeroScience in 2022. Biological age is an attractive shortcut for trials, they wrote, but it only earns that role once studies show that a treatment's effect on the score reliably predicts its effect on real health outcomes. That proof does not exist yet for any clock.
What can a consumer test tell me, and what can't it?
A result from a well-validated clock is a risk signal. People who score older on GrimAge or faster on DunedinPACE are, on average, more likely to become ill and die sooner. That is real science, but much of the signal reflects risks that cheaper tests already measure. GrimAge is partly built from a methylation stand-in for smoking history, and in the 2019 paper it also tracked excess fat in the liver and around the organs on scans.
The free NHS Health Check, offered every 5 years to people aged 40 to 74 without existing heart, kidney or blood sugar conditions, measures weight, waist, blood pressure and cholesterol, asks about smoking, drinking and activity, and gives you a 10-year risk score for heart disease, stroke, type 2 diabetes and kidney disease. Doctors know how to act on those numbers. They do not yet know how to act on an epigenetic age.
What a consumer test cannot tell you is how long you will live, whether a particular supplement or routine is working for you, or whether a change of a year or two is real. A claim that a product reversed someone's biological age is a claim about a score, not about health, unless it comes from a controlled trial that also measured how people actually fared.
This lesson is general education about biological-age tests, not medical advice. If a result worries you, or you want to understand your own health risks, talk to your GP.
Related in this classroom Related in other classrooms The studies behind this page DNA methylation age of human tissues and cell types - Genome Biology, 2013An epigenetic biomarker of aging for lifespan and healthspan - Aging (Albany NY), 2018DNA methylation GrimAge strongly predicts lifespan and healthspan - Aging (Albany NY), 2019DunedinPACE, a DNA methylation biomarker of the pace of aging - eLife, 2022A computational solution for bolstering reliability of epigenetic clocks: implications for clinical trials and longitudinal tracking - Nature Aging, 2022Effect of long-term caloric restriction on DNA methylation measures of biological aging in healthy adults from the CALERIE trial - Nature Aging, 2023Endpoints for geroscience clinical trials: health outcomes, biomarkers, and biologic age - GeroScience, 2022NHS Health Check - NHSFurther reading
Sources for this classroom
Educational content only. Not medical advice, diagnosis or treatment. Always consult a qualified healthcare professional before changing your health regimen.