Your chronological age is just the candles on the cake. Your biological age — how old your body actually is at the cellular level — can be higher or lower, and DNA-based tests now attempt to measure it by reading the methylation patterns on your genes. These “epigenetic clocks” have advanced a great deal, and the newest ones can estimate not just how old you are biologically but how fast you’re currently ageing. They’re genuinely fascinating and increasingly useful — but they come with real caveats, and a single number from one test means far less than the marketing suggests. The value is in the trend and the context, not the headline figure.
Chronological age versus biological age
Everyone the same age is not ageing at the same rate. Two 50-year-olds can have bodies that function like a 40-year-old’s or a 60-year-old’s, depending on genetics, lifestyle, stress, illness and exposures accumulated over decades.
Chronological age is fixed — one year per year, nothing you can do. Biological age is the more interesting number, because it reflects the wear and repair actually happening in your cells, and unlike your birthday, it can potentially move in either direction. The question that’s driven a whole field of research is: can we measure it?
How epigenetic clocks work
The breakthrough came from epigenetics — specifically DNA methylation, the chemical tags that sit on your DNA and switch genes on or off without changing the underlying code.
These methylation patterns change in predictable ways as we age. In 2013, Steve Horvath showed that by measuring methylation at a set of specific sites across the genome, you could estimate a person’s age remarkably accurately — the first “epigenetic clock”¹. Crucially, because methylation reflects gene expression rather than the fixed genetic blueprint, it’s influenced by how you’ve lived — which is what makes it potentially responsive to diet and lifestyle, rather than a fixed readout.
That original insight has since developed into several generations of clocks, and the differences between them matter.
The generations of clocks — and why the newest ones are better
This is the part that’s changed most, and it’s worth understanding because not all “biological age” tests measure the same thing or mean the same thing.
First generation (Horvath, Hannum). Trained to predict chronological age from methylation. Impressive technically, but their goal was matching your actual age — so they’re less informative about health and disease risk².
Second generation (PhenoAge, GrimAge). A significant step forward. Instead of being trained on age alone, these were trained on clinical health markers and mortality. GrimAge in particular incorporates methylation surrogates for things like smoking and inflammation, and predicts all-cause mortality well enough to have earned the nickname “the death clock”³. GrimAge version 2 improved this further by adding markers for inflammation (CRP) and blood sugar (HbA1c)³. These clocks correlate with real outcomes — physical function, cognitive decline, frailty — not just age³.
Pace-of-aging measures (DunedinPACE). The newest and conceptually different. Rather than estimating how old you are cumulatively, DunedinPACE estimates how fast you are ageing right now — your rate, not your total⁴. It has high test-retest reliability and is strongly associated with morbidity, disability and mortality⁴. For tracking whether an intervention is actually slowing your ageing, a pace measure is more useful than a cumulative one, because it responds to your current state.
The practical upshot: if you’re looking at a biological-age test, which clock it uses matters. A test naming GrimAge2, PhenoAge or DunedinPACE is measuring something validated against health outcomes. A test giving you a single “age” from an older or unnamed clock is telling you much less⁵.
The honest caveats — and they’re significant
This is where most consumer coverage of biological-age testing falls short, so it deserves real space. These tests are promising, not oracular.
A single reading is unreliable. Epigenetic age has been shown to fluctuate by as much as two years over the course of a single day, and it’s affected by fasting status, acute stress, menstrual cycle and even time of year⁶. So one measurement, taken once, can be off by more than the difference you’re hoping to detect. This is a real limitation, not a technicality.
One clock isn’t enough. Because of that variability, researchers increasingly argue that a genuine change should show up across several clocks, not just one — and that “principal component” (PC) versions of the clocks are more reliable for tracking change over time⁵ ⁶. A test that reports one number from one clock is the shakiest version of this.
“Reversing your age” is contested. You’ll see dramatic claims about reversing biological age through diet and lifestyle. The most-cited study (Fahy and colleagues, 2019) reported an apparent reversal — but it was tiny (around 38 completers), men only, and used an older clock⁷. Later analysis has argued that much of what looks like “reversal” in such studies may be regression to the mean — the statistical tendency for extreme readings to drift back toward average on retesting⁸. So the idea that you can meaningfully wind your biological clock backwards remains genuinely unsettled, however appealing.
None of this means the tests are worthless. It means they’re best read as one input, ideally tracked over time with a reliable clock, as part of a bigger picture — not as a precise verdict you can act on from a single result. Which, as with most testing, is exactly the right way to use them.
What genuinely does accelerate and slow biological ageing
Here’s the encouraging part, and unlike in 2019 it’s now backed by large longitudinal studies rather than inference. A 2025 multi-cohort study using tools like DunedinPACE found clear patterns⁹:
Accelerates ageing: smoking, higher BMI, elevated blood glucose, and poor blood pressure profiles — the markers of metabolic and cardiovascular strain, with inflammation running through much of it⁹.
Slows ageing: regular physical activity and a healthier diet, with favourable cardiovascular health consistently linked to slower epigenetic ageing⁹.
Translated into the things you can actually influence:
- Reduce the inflammatory load — since chronic inflammation underlies much of accelerated ageing. That means addressing the usual drivers: poor diet, unresolved infections, chronic stress, toxin exposure.
- Eat a diet rich in plants and polyphenols, which support healthy cellular function (and, as covered elsewhere, work partly through the gut microbiome).
- Keep blood sugar and metabolic health in good order — glucose regulation shows up repeatedly as an ageing lever.
- Move regularly, but don’t overtrain — moderate activity slows ageing; chronic overtraining is a stressor that can push the other way.
- Support the basics — sleep and circadian rhythm, stress management, nutritional adequacy for methylation. These aren’t glamorous, but they’re what the data keeps pointing to.
Notice that none of this is exotic. The interventions that move biological age are, overwhelmingly, the fundamentals of metabolic and inflammatory health — which is either disappointing or reassuring, depending on how you look at it.
Where testing fits
Biological-age testing can be a genuinely useful motivator and tracker — a way to see whether the changes you’re making are registering, retested over a meaningful interval (many recommend 6–12 months) rather than chased month to month. Used that way, with a reliable clock and realistic expectations, it turns abstract “healthy living” into something measurable.
Just hold the number lightly. It’s a snapshot of a noisy, fluctuating measure, most meaningful as a trend over time and alongside the rest of your health picture — not a single score to fixate on. A biological-age result, like any marker, means little stripped of its context: which clock produced it, under what conditions, and which way it’s moving over time.
The short version
DNA methylation-based epigenetic clocks can now estimate your biological age, and the newest ones (like DunedinPACE) can estimate how fast you’re ageing right now. The science is real and advancing, and the lifestyle factors that move biological age — managing inflammation, metabolic health, activity, diet — are now well evidenced. But a single reading is genuinely unreliable, “reversing” ageing is still contested, and which clock a test uses matters enormously. Treat the number as a trend to track in context, not a verdict.
Frequently asked questions
What is an epigenetic clock?
A test that estimates your biological age by measuring DNA methylation — the chemical tags on your genes that change in predictable ways as you age. Because methylation reflects gene expression rather than your fixed genetic code, it’s influenced by how you’ve lived, which is what makes biological age potentially changeable.
What’s the difference between biological age and chronological age?
Chronological age is simply how many years you’ve lived and can’t change. Biological age reflects the actual wear and repair happening in your cells, can be higher or lower than your chronological age, and can potentially move in either direction depending on health and lifestyle.
Which epigenetic clock is best?
It depends what you want to know. GrimAge (and GrimAge2) are strongest for predicting mortality and health risk; DunedinPACE is best for measuring your current rate of ageing; first-generation clocks like Horvath’s are better at estimating chronological age than health. A test that names its clock is more trustworthy than one that doesn’t.
Are biological age tests accurate?
They’re promising but come with real caveats. Epigenetic age can fluctuate by up to two years within a single day and is affected by fasting, stress and other factors, so a single reading is unreliable. Using a reliable clock, tracking over time, and not over-interpreting one number all matter.
Can you reverse your biological age?
This is genuinely contested. Some studies suggest lifestyle changes can lower biological age, but the most-cited one was very small, and later analysis argues much of the apparent “reversal” may be regression to the mean. Slowing the rate of ageing is better supported than reversing it.
What slows biological ageing?
Large recent studies point to the fundamentals: not smoking, healthy body weight, good blood sugar and blood pressure, regular moderate physical activity, and a healthy plant-rich diet. Chronic inflammation runs through most of the accelerators, so reducing inflammatory load is a common thread.
References
- Horvath, S. (2013). DNA methylation age of human tissues and cell types. Genome Biology, 14, R115 — the first widely used epigenetic clock.
- First-generation clocks (Horvath, Hannum) were trained to predict chronological age and are less informative about health outcomes than later clocks. Epigenetic clock reviews (2025), Frontiers in Molecular Biosciences; Biogerontology (2026).
- Second-generation clocks PhenoAge and GrimAge are trained on clinical phenotypes and mortality; GrimAge predicts all-cause mortality strongly, and GrimAge2 adds CRP and HbA1c surrogates, improving risk stratification. Liang et al. (2025), PMC12539533.
- Belsky, D. W. et al. (2022). DunedinPACE, a DNA methylation biomarker of the pace of aging. eLife, 11:e73420. High test-retest reliability; measures rate rather than cumulative age.
- Tests naming validated clocks (GrimAge2, PhenoAge, DunedinPACE) and using principal-component (PC) reformulations are more reliable; single-number results from unnamed or older clocks tell you much less. Epigenetic age testing guide (2026); Higgins-Chen et al. (2022), Nature Aging (PC-clocks).
- Epigenetic age can fluctuate by up to ~2 years across a single day and varies with fasting, acute stress, menstrual cycle and time of year; multiple reliable clocks should agree before a change is trusted. “When to Trust Epigenetic Clocks,” PMC11526921; Koncevičius et al. (2024).
- Fitzgerald/Fahy-era reversal claims: the most-cited intervention studies were small (e.g. ~38 completers, men only, older saliva-based clock), limiting confidence. Epigenetic age testing guide (2026); Fahy, G. M. et al. (2019), Aging Cell, 18(6), e13028.
- A 2024 analysis argues much of the apparent epigenetic-age “reversal” in intervention studies may reflect regression to the mean rather than true rejuvenation. Yale bioRxiv preprint (2024), summarised in epigenetic testing reviews (2026).
- Longitudinal multi-cohort data (2025): smoking, higher BMI, elevated glucose and poor blood pressure accelerate epigenetic ageing (DunedinPACE), while physical activity and healthier diet slow it; favourable cardiovascular health links to slower ageing. eBioMedicine / BMC Medicine (2025), summarised in PMC12905613.
