What the COSMOS study in Nature Medicine really says about multivitamins and epigenetic clocks

A closer read of the paper, however, paints a far more sober picture.

Across many health columns it suddenly sounds like a simple fix: take one pill a day and your body ages more slowly. The spark for this is a recent study in Nature Medicine that followed adults aged over 70. The project, called COSMOS, asks whether a daily multivitamin could slow the body’s so‑called “epigenetic clockwork” - in other words, biological ageing processes measured at DNA level.

What the COSMOS study is about

The COSMOS study (“Cocoa Supplement and Multivitamin Outcomes Study”) comes from a well‑known research group at Brigham and Women’s Hospital in Boston. The idea was to test not individual vitamins, but a standard over‑the‑counter combination supplement under real‑world conditions.

Just under 1,000 people took part, with men and women represented equally and all participants aged at least 70. They were randomly allocated to one of four groups and took capsules every day for two years - without knowing what they were actually receiving.

Four groups, four different treatments

• Multivitamin supplement plus 500 mg cocoa powder (with about 50 mg epicatechin, an antioxidant)
• Cocoa powder only plus placebo
• Multivitamin supplement only plus placebo
• Pure placebo with no active ingredient

This design was intended to show whether cocoa components, the multivitamin, or the combination produced measurable changes in biological markers linked to ageing.

How ageing was measured in the lab

Rather than counting wrinkles or running fitness checks, COSMOS centred on a more abstract laboratory metric: so‑called epigenetic clocks. These refer to patterns of chemical “tags” on genetic material - especially methyl groups that attach to or detach from DNA. Such patterns shift with age and are treated as indicators of “biological age”.

In practice, the researchers measured five epigenetic signatures in all participants. Blood samples were analysed at the start of the trial, after 12 months, and again after 24 months. Two of these clocks are considered particularly closely associated with mortality.

"The analysis shows that people who took a multivitamin supplement had slightly slowed changes across all five epigenetic clocks - especially among those who appeared to be ageing faster at baseline."

Across the two‑year period, the model translates this into an advantage of roughly four months compared with the placebo group. Put another way: after two calendar years, cells in the multivitamin group looked, on average, as though only just under 20 months had passed.

Cocoa showed no measurable anti‑ageing effect

The study’s other major candidate - cocoa powder containing epicatechin - did not produce benefits in the epigenetic clocks. No improvements were detected either on its own or alongside the multivitamin. That leaves the spotlight on the vitamin supplement, even if the measured effect is modest.

What the researchers themselves say the findings mean

Despite the headlines, the authors stress restraint. In their conclusions they explicitly describe the results as “statistically significant, but limited”. Whether such differences would be noticeable in day‑to‑day life remains unanswered.

"The researchers emphasise that further studies are needed to clarify whether a slightly altered epigenetic age really translates into fewer diseases and a longer healthy lifespan."

One noteworthy point comes from lead investigator Howard Sesso in an accompanying piece in Nature: he argues that the interaction between diet, nutrient intake and epigenetic clocks needs to be examined more broadly. The frank reality is that nobody currently knows whether a multivitamin has the same effect as - or a weaker effect than - consistently eating a genuinely healthy diet.

That comparison is precisely what COSMOS did not include. The trial never set up “vitamin pill versus Mediterranean‑style eating” or “vitamin pill versus substantially more vegetables, pulses and wholegrains”. Instead, it simply compared a pill with a placebo against a background of fairly typical everyday diets in the United States.

How reliable are epigenetic clocks?

Epigenetic clocks are widely seen as an exciting area of research. Some models can estimate biological age with surprising accuracy and even predict mortality risk in statistical terms. Even so, the approach is not yet established in routine clinical practice.

• The testing methods are complicated and expensive.
• There are many different clocks, each using different algorithms.
• The direct link to specific diseases is still poorly understood.
• It remains unclear how strongly lifestyle changes shift these values in a lasting way.

Sesso also notes that plenty of pieces are still missing: the exact relationship between epigenetic patterns and chronic age‑related diseases such as heart attack, dementia or cancer is only loosely mapped out. The same uncertainty applies to how diet, exercise or supplements influence these markers.

Who funded the study?

To interpret medical research sensibly, it is worth checking the end of the paper - where funders and potential conflicts of interest are listed. For COSMOS, the backers form a mixed group: public funders alongside major organisations from the nutrition and supplement industries.

They include, among others:

• The US National Institutes of Health (NIH)
• Mars (including its health division, Mars Edge, linked to cocoa products)
• Pfizer Consumer Healthcare (now Haleon), a major manufacturer of over‑the‑counter products
• Foxo Technologies, which specialises in epigenetic testing
• American Pistachio Growers
• Council for Responsible Nutrition, a trade association for the supplement industry

The authors state they worked independently and that there are no financial conflicts of interest. Formally, that meets common transparency standards. In public debate, though, such a strong mix of industry‑adjacent funding can still leave an unpleasant aftertaste.

If identical results had emerged without support from large supplement manufacturers, the findings would be easier to assess calmly. Where effects are relatively small, questions about possible bias naturally arise - even if no bias can be demonstrated.

What does this mean for older people in everyday life?

For older adults, the message sounds tempting: an inexpensive multivitamin from a high‑street shop and biological ageing slows down. In reality, it is more complicated.

Aspect	What the study indicates
Epigenetic age	Slightly slowed ageing markers with multivitamin use
Duration	Two years of follow‑up in adults aged over 70
Clinical outcomes	No clear evidence of fewer heart attacks, cancers or dementia
Comparison with diet	No direct head‑to‑head with very healthy eating
Funding	Money from public bodies and several industry stakeholders

For some individuals, multivitamins can be appropriate - for example, where deficiencies are confirmed, appetite is poor, diets are limited, or certain medication combinations are involved. Many GPs now check vitamin D, B12 or folate specifically.

The COSMOS data add one more argument that a daily tablet might offer a small biological advantage for very old people whose diets are less than ideal. They do not, however, amount to a free pass to disregard other lifestyle factors.

Vitamins from a pill or from the plate?

Specialists in nutritional medicine have been repeating the same message for years: most nutrients should come from food. A whole‑food diet delivers not only vitamins, but also fibre, plant compounds and complex combinations that cannot yet be fully replicated in a tablet.

In everyday terms, a realistic anti‑ageing package usually involves:

• lots of fruit and vegetables, with as much variety and colour as possible
• wholegrains instead of refined white flour products
• pulses, nuts and seeds
• high‑quality plant oils and oily fish
• regular exercise, particularly strength training plus moderate endurance work
• not smoking and keeping heavily processed convenience foods to a minimum

A multivitamin may plug a gap in certain situations, but it cannot replace that wider package. A compelling future trial would tackle the question directly: how does a multivitamin supplement perform in a straight comparison with a dramatically improved diet, perhaps paired with structured training?

Until then, the Nature Medicine paper is mainly one thing: an intriguing piece of a much larger ageing puzzle. It suggests that micronutrients can leave measurable traces in our cells. How much those traces ultimately determine health, disease and lifespan is something research still needs to establish - ideally with independent funding and an honest comparison with the old‑fashioned vegetable stir‑fry.