In a six-month exercise program, more training was linked to slower-ticking epigenetic clocks

Your birthday tells you how old you are, but not how fast your body is aging. To estimate that, researchers use epigenetic clocks. Along our DNA sit small chemical tags called methyl groups. They don't change the genetic code, but they help switch genes on and off, and at certain spots on the DNA their pattern shifts steadily with age. An epigenetic clock reads a selected set of these spots in a blood sample and turns them into an estimated age. Newer clocks such as PhenoAge and GrimAge were built to track health and the risk of disease and death, not just years lived.

If a clock says you are older than your real age, that difference is called age acceleration. A negative number means the clock reads younger. Earlier work found that fitter people tend to have lower age acceleration, but there are few studies that follow people through an exercise program, and nobody knew how much exercise it takes to move these clocks. A team funded by Hungarian research agencies set out to measure that "dose". Their study is a preprint: it has not yet been peer-reviewed.

What they did

The team enrolled 78 healthy volunteers aged 50 to 70, 36 men and 42 women. After six get-to-know-the-machine sessions, they trained for six months on stationary bikes, elliptical trainers and rowing machines, doing a form of interval training (alternating harder and easier stretches) that got gradually more demanding.

Sessions were offered three times a week, and coming was voluntary, so people ended up with very different amounts: from 15 to 74 sessions each. Four people dropped out and two more were left out for attending fewer than 35 sessions. That spread let the team compare light and heavy trainers.

They expressed each person's total exercise in MET-hours. A MET (metabolic equivalent) compares an activity's energy use with sitting still; one hour at an intensity of 5 METs counts as 5 MET-hours. Before and after the program, they measured several epigenetic clocks, scanned the heart with ultrasound (echocardiography), measured fat and muscle with a body scan (DEXA), took blood tests and, for 52 people, analysed gut bacteria from stool samples.

What they found

• More exercise, bigger drop on two clocks. Higher MET-hours were linked to a larger fall in age acceleration on versions of PhenoAge and GrimAge. In the team's simplest model, each extra 110 MET-hours or so over six months went with about one year less PhenoAge acceleration; for GrimAge it was about 202 MET-hours.
• A rough "dose". The model predicts that someone who does no exercise would drift toward a positive age acceleration. To end up one year below zero from that starting point, it estimates about 255 MET-hours over six months for PhenoAge and 378 for GrimAge. The team calls this broadly comparable to the World Health Organization's activity guidelines, which they put at roughly 260 to 520 MET-hours per six months.
• Not every clock agreed. Across the whole group, three clocks (including PhenoAge) read younger after the program. But a methylation-based estimate of telomere length (the protective caps on chromosomes) went down, and DunedinPACE, a clock that measures the speed of aging, went up.
• Other changes. On average, participants lost body weight and fat, and gained leg muscle. HDL ("good") cholesterol rose; diastolic blood pressure, fasting blood sugar and total cholesterol fell. Estimated fitness, resting heart rate and heart-rate recovery improved; grip strength did not. A short-term memory test (repeating digits) improved.
• The heart. A measure of heart-wall thickness decreased, and people whose wall thickness fell more also tended to show a larger drop in PhenoAge acceleration.
• Gut bacteria became more diverse. Two blood molecules linked to gut bacteria and exercise, TMAO and Lac-Phe, changed in ways whose meaning the team says is still unclear.

What to keep in mind

• No control group. The authors write that without a group that didn't exercise, the before–after changes can't be taken as proof of cause, and that randomized trials are needed.
• People chose their own dose. Those who came more often may differ from those who came less, in health, motivation or daily habits outside the gym. The dose–response link is an association.
• Small, model-based numbers. The "255 MET-hours" figure comes from a fitted line in about 70 people; the team expects the real relationship is not a straight line.
• Mixed clock results. DunedinPACE and the telomere estimate moved the "wrong" way, a reminder that different clocks measure different things.
• Heart findings are tentative. The authors call the link between clocks and heart shape "speculative", and one heart-filling result only "exploratory". Another heart measure showed only a trend (p = 0.052).
• The participants were healthy volunteers, so the results may not apply to people with heart or other diseases. This study is not advice on how much or how hard to exercise.

Why it matters

This study tries to put a number on how much exercise goes with a measurable change in epigenetic clocks, and it found that clock changes moved together with changes in the heart and fitness. That hints the clocks pick up something real in the body, though the team says this needs confirming. For now, it is a small, uncontrolled preprint that still has to pass peer review, and randomized trials will have to show whether the pattern holds.