Do multivitamins slow biological aging? What the COSMOS trial found

A large US trial found a daily multivitamin slowed two aging clocks a little. Those are markers, not proof that you will live longer.

Published October 2026

Short answer

Early, limited signs that a daily multivitamin slows some aging markers. Not proof of a longer life, and the big heart trials found no benefit. Evidence grade C (Limited). One prespecified two-year trial substudy (958 people) plus a small metabolomic trial; downgraded once because the outcome is a surrogate aging marker.

Key takeaways

In a two-year part of the COSMOS trial, a daily multivitamin slowed the rise of two of five DNA-based aging clocks compared with placebo [1].
The slowing was larger in people whose biological age was already running fast at the start of the study [1].
A blood-metabolite analysis of 399 people found small rises in omega-3 and DHA, but its metabolic aging scores were not statistically significant [2].
The authors say the clinical meaning of the clock changes is still unknown [1].
Large trials of heart disease found no benefit, and cancer results are mixed [3][4][5].

The short version

A daily multivitamin slowed some measures of biological aging in a large US trial of older adults, but only a little. In a two-year part of the COSMOS trial, two of five DNA-based aging clocks rose more slowly on a daily multivitamin than on placebo [1]. These are markers, not proof of a longer life, and the evidence is limited. The large trials of heart disease found no benefit [3][4].

What the studies show

COSMOS, the COcoa Supplement and Multivitamin Outcomes Study, randomly assigned older US adults to a daily multivitamin-mineral pill, cocoa extract, both, or neither [4].

The aging report used DNA methylation, chemical marks on DNA, to score five aging clocks in 958 participants over two years [1]. The multivitamin slowed the rise of two of them. PCGrimAge rose 0.11 years less per year than on placebo, and PCPhenoAge rose 0.21 years less per year [1]. Among people whose biological age was already running fast, the PCGrimAge effect was about 0.24 years per year [1]. The cocoa extract had no effect on any of the clocks [1].

A second report on 399 participants measured about 170 blood metabolites [2]. The multivitamin slightly raised omega-3 fatty acids and docosahexaenoic acid (DHA) and slowed a rise in creatinine, a blood waste marker [2]. Its metabolic aging clocks dropped by about 0.5 to 1 year, but the change was not statistically significant [2].

The nuances

Clock scores are markers, not diseases. The authors say the clinical relevance of the change is unknown and that more studies are needed [1].

Hard outcomes tell a different story. In 14,641 male physicians followed for about 11 years, a daily multivitamin did not reduce major cardiovascular events [3]. In COSMOS-Web, over 3.6 years, heart events were similar on both groups, and deaths from any cause were not significantly lower [4].

Cancer results are mixed. A trial in male physicians found a small reduction in total cancer [5], but COSMOS-Web found no significant reduction [4]. For memory, a COSMOS trial of 2,262 people found small gains in global cognition after three years [6], while the Physicians' Health Study II found no effect in men aged 65 and older [7].

The aging report's authors disclosed grants from Mars Edge, and Mars Edge and Pfizer Consumer Healthcare (now Haleon) donated the study pills and packaging [1].

What this means for you

Multivitamins are widely used, with at least one-third of US adults taking one [5]. The aging findings are interesting, but they do not show that a multivitamin will slow aging in you [1]. Talk with a doctor or pharmacist before you start a supplement, especially if you take medicines or have kidney disease. The multivitamins topic page lists the full set of studies and safety points.

Go deeper

Multivitamins — A daily multivitamin may slightly improve memory in older adults and slowed two DNA-based aging clocks in one trial, but the evidence for both is limited and m…