Omega-3 slowed biological-age clocks in a trial, by a few months over three years

A daily gram of omega-3 modestly slowed several DNA-methylation "clocks" of biological ageing in a randomised trial of older adults, the first such evidence from a proper controlled study rather than a supplement company's marketing [s1]. The effect was very small, equivalent to a few months of slower ageing over three years, and it came from the same trial that had already found the supplements did nothing for blood pressure, cognition, infections or fractures [s1][s2].

What was measured

The analysis draws on DO-HEALTH, a European trial that randomly assigned generally healthy adults aged 70 or older to combinations of vitamin D at 2,000 international units a day, omega-3 at 1 gram a day, and a home strength-exercise programme, in a factorial design that let researchers separate each treatment's effect [s2]. In a pre-planned but post-hoc sub-study, investigators measured biological ageing in 777 of those participants using four "next-generation" epigenetic clocks, tools that estimate biological age from chemical methylation marks on DNA: PhenoAge, GrimAge, GrimAge2 and DunedinPACE [s1].

Over three years, omega-3 alone slowed three of the clocks, PhenoAge, GrimAge2 and DunedinPACE, and all three interventions together had an additive benefit on PhenoAge [s1]. The size of the effect is the part that deserves emphasis. Across the clocks, the standardised effects ranged from 0.16 to 0.32 units, which the authors translate to roughly 2.9 to 3.8 months of slower biological ageing over the three-year study [s1]. That is a real, statistically detectable signal, and it is also, in absolute terms, tiny.

Why the parent trial matters

The reason to read this result cautiously is sitting in the same dataset. DO-HEALTH's main report, published in 2020, tested whether these supplements improved six clinical outcomes in 2,157 adults over three years: systolic and diastolic blood pressure, physical performance, cognition, and rates of non-vertebral fractures and infections [s2]. The participants had a mean age of 74.9 years and were 61.7% women [s2]. The trial found no significant benefit on any of them [s2].

So the honest summary is uncomfortable for the supplement industry: the same trial that moved a biological-age clock by a few months moved none of the health outcomes those clocks are supposed to predict. That gap is the whole problem with treating epigenetic clocks as if they were the goal. A clock is a surrogate, a stand-in for ageing that is easy to measure, and a change in a surrogate is only as meaningful as its link to something a person can feel, a fracture avoided, a year of independent living, a death postponed. DO-HEALTH shows the surrogate moving while the real outcomes stayed put.

There are further limits. The clock analysis was post-hoc, meaning it was not the trial's original purpose, which weakens it relative to a pre-registered primary result [s1]. The clocks themselves are an unsettled measurement, and their reliability and reproducibility are actively debated; a small shift in a noisy instrument is exactly the kind of finding that needs replication before it is believed. And the participants were healthy, older and mostly European, so the result does not automatically transfer to younger or sicker populations [s2].

What it means for a reader

This is a genuinely interesting finding and a poor basis for buying fish-oil capsules. The strongest evidence for omega-3 remains the large cardiovascular and cognitive trials, which have been mixed and largely disappointing for healthy people, and the same "moves a marker, not an outcome" pattern dogs vitamin D in healthy adults. A few months on a methylation clock does not overturn that.

It is also a caution about the consumer market this feeds. Direct-to-consumer biological-age tests are sold on the premise that you can measure your ageing and then buy products to reverse it, and this trial is already being cited to that end. What DO-HEALTH actually supports is narrower: that omega-3 can nudge a research clock by a small amount, in a study whose clinical bottom line was null [s1][s2]. Whether epigenetic clocks are trustworthy enough to serve as endpoints in ageing trials at all is still an open scientific question, not a settled marketing claim.

What to watch next is whether purpose-built trials, ones that set out to change a clock and then follow people long enough to see whether their health actually differs, can close the gap between the surrogate and the outcome. Until then, the accurate reading of the first randomised evidence that a supplement slows biological ageing is that the effect is real, small, and so far disconnected from anything a person would notice.