A Popular Ageing Test Can Move, but That Doesn’t Mean You’ll Live Longer

You may have seen the ads. Send off a blood sample, and a lab tells you your “biological age”.

Some people then try diets, supplements or drugs, and test again to see if the number drops.

But do these “ageing clocks” really respond to anything? And if they move, which ones should we trust?

A new study in Nature Medicine is the most systematic attempt yet to find out. It pooled 51 human studies that measured the clocks before and after an intervention.

The short answer: some clocks respond, some barely do, and drugs moved them more than lifestyle changes.

What is an ageing clock?

Chemical tags on your DNA

Your DNA carries tiny chemical tags called methyl groups. The pattern of these tags shifts as you age.

Scientists have trained computer models to read those patterns in a blood sample. These models are called epigenetic clocks.

Different kinds of clock

Not all clocks measure the same thing.

The newer clocks were built to track health, not just birthdays. That turns out to matter a lot.

Why researchers want them

Testing whether a treatment slows ageing normally means waiting decades to see who gets ill or dies. A reliable clock could give an early signal within months.

But first, a clock has to respond when something genuinely works. Surprisingly, that had never been tested across many studies at once.

The study: 51 interventions, 16 clocks

Pulling the evidence together

Researchers led by a team at Yale built a database called TranslAGE. It gathered 51 studies with blood samples taken before and after an intervention.

Together, the studies held 3,128 samples. For every study, the team calculated the same 16 main clocks, plus 94 other DNA-based markers.

What kinds of interventions?

The interventions fell into four groups:

• Lifestyle: diets and exercise, including Mediterranean, vegan and low-carb diets.
• Drugs: such as metformin, rapamycin, semaglutide and anti-inflammatory anti-TNF drugs.
• Supplements: such as omega-3 and folate.
• Medical procedures: such as hyperbaric oxygen therapy, organ transplants and gene therapy.

Most studies (41) involved healthy people. The other 10 involved people with a disease.

Checking the method works

First, the team checked that the approach could tell good events from bad ones.

They ran the same analysis on control groups and on nine “pro-ageing” events. Those events pushed the clocks up, while the interventions on average pushed them down.

A fair comparison

Each study’s results were adjusted for people’s ages and put on the same scale. That let the team compare very different interventions side by side.

For each person, they compared the clock reading before and after the intervention.

The results: which clocks respond?

Overall

Across the 16 main clocks, 19 of the 51 interventions lowered epigenetic age. Thirteen of those held up after a stricter statistical check.

Five interventions raised epigenetic age, and 26 showed no clear effect.

Newer clocks respond best

The clearest winners were the newer, more reliable versions of the second-generation clocks.

DunedinPACE showed the biggest average drop. It fell significantly in 16 interventions and rose in only one.

PCGrimAge gave the most statistically solid results. GrimAgeV2, PCPhenoAge and SystemsAge also fell reliably across interventions.

The original first-generation clocks, which mostly track calendar age, moved up and down with no clear pattern.

Drugs moved the clocks most

Anti-TNF drugs, used for arthritis and bowel disease, stood out. They lowered almost every second-generation clock, and did so again in a second study.

Two different Mediterranean diets also lowered similar sets of clocks in healthy people. That consistency is what makes a result believable.

Sick people showed bigger changes

The clocks moved more in people with a disease than in healthy volunteers.

That makes sense. There is more room for improvement when the body is under strain.

DunedinPACE was the exception. It responded about equally in both groups, which may make it the most dependable across different people.

Clocks can disagree

Not every intervention gave a tidy picture.

Five studies of senolytics, drugs meant to clear out old “zombie” cells, gave mixed results. Some clocks went up while others went down, sometimes in the same study.

That’s a warning sign. When clocks disagree, it’s hard to know what the treatment really did.

The authors propose two simple rules for a believable result. Similar clocks should move in the same direction, and a second study of the same intervention should show the same thing.

Anti-TNF drugs and Mediterranean diets passed both tests. Senolytics did not.

Signal versus noise

Why do the older clocks do so badly? The team found that much of their movement looks like random noise in the DNA tags.

Interventions seem to change the more structured, biological part of the signal. The newer clocks capture more of that part.

How big were the changes?

The average changes look small on paper. But the authors argue they aren’t trivial.

For clocks such as GrimAge and PhenoAge, a standard effect of about 0.2 corresponds to roughly one year of epigenetic age.

A closer look at body systems

The “explainable” clocks split ageing into different body systems. They picked up effects the single-number clocks missed.

Stopping smoking mainly lowered the lung score, for example. Metformin moved inflammation, brain and metabolic scores the most.

Across seven different diets, one score fell every time: the musculoskeletal system, meaning muscles and bones.

What earlier research found

A calorie restriction trial

In the CALERIE trial, 220 healthy adults without obesity were randomly assigned to cut calories by 25% or eat normally for two years.

A 2023 analysis in Nature Aging found calorie restriction slowed DunedinPACE. It didn’t change PhenoAge or GrimAge, and the effects were small.

Omega-3, vitamin D and exercise

The DO-HEALTH trial tested vitamin D, omega-3 and home exercise in 777 older adults over three years.

A 2025 analysis found omega-3 slowed three clocks, including DunedinPACE. The effects were worth roughly three to four months of biological ageing.

Everyday exposures

A 2026 meta-analysis pooled 83 mostly observational studies. Unhealthy exposures, such as poor metabolic health and air pollution, were linked to faster epigenetic ageing.

How it fits together

Randomised trials suggest real interventions can nudge the newer clocks, but only a little. The new study fits that pattern across many more interventions.

It also helps explain why earlier results often looked inconsistent. Studies used different clocks, and some clocks are much noisier than others.

How much should you trust this?

Early. It’s a big, careful comparison, but it tells us which clocks move, not whether moving them helps you live longer.

What makes it convincing

• It applies the same 16 clocks to every study, so results can be compared fairly.
• It checks for consistency across clocks and across repeat studies.
• Pro-ageing events and control groups behaved as expected.
• It was published in a leading journal after peer review.

What makes me cautious

• The 51 studies differed hugely in size, length, people and design.
• Study quality and size varied, which the authors acknowledge.
• Nobody knows yet how much change in a clock is enough to matter for health.
• Some authors work for, or consult for, a company that sells epigenetic tests.
• Some clocks include inflammation markers, so they may track inflammation rather than ageing itself.

What this means for you

If you’re thinking of buying a biological age test, this study is a useful reality check.

• Ask which clock it uses. A result from an older, calendar-age clock tells you little about your health.
• Don’t chase small changes. Clocks can be noisy, and different clocks can give different answers.
• Be wary of supplement claims. Supplements as a group didn’t move the clocks here.
• Stick with the basics. Diet and exercise nudged the clocks, and they have plenty of other proven benefits.

The World Health Organization explains what shapes healthy ageing, from lifestyle to surroundings.

In this TEDx talk, Steve Horvath, who built one of the first epigenetic clocks, explains how they could help find anti-ageing treatments:

What we still don’t know

1. Does lowering a clock mean longer life? Only long-term trials that track disease and death can say.
2. How big a change matters? There’s no agreed “meaningful” drop yet.
3. Which clock should trials use? DunedinPACE and GrimAge-type clocks look best so far.
4. Do effects last? Longer follow-up is needed to see whether changes stick.
5. Does it work in healthy people? The clearest effects appeared in people with disease.
6. What do explainable clocks add? System-level scores look promising but need more testing.

My take: the ruler matters as much as the treatment

What I like about this study is that it tests the ruler, not just the treatment. Before we ask “does this slow ageing?”, we need clocks that respond in a sensible way.

The good news is that some do. DunedinPACE and the GrimAge family behaved consistently across very different interventions.

I’m less convinced by consumer tests that report a single “biological age”. This study shows how much the answer can depend on which clock you use.

For now, I’d treat these clocks as promising research tools. They’re not yet a scorecard for your own health.