# Daily multivitamin (COSMOS)

Rating: Promising — Some human trials, or consistent animal work with a clear mechanism
Type: Compound
Updated: 2026-10-18
Canonical: https://heretic.bio/articles/daily-multivitamin-cosmos-trial.html

> A daily multivitamin has unexpectedly good randomized evidence for slowing cognitive aging in older adults. In COSMOS, a pooled analysis of three cognitive substudies found the multivitamin slowed memory and global cognitive decline by the equivalent of about two years, and a 2026 analysis reported slower epigenetic aging. The same trial found no effect on cancer or cardiovascular disease, and the earlier Physicians Health Study II found no cognitive benefit in male doctors.

For two decades the standard line has been that multivitamins are expensive urine: large trials found no protection against heart disease or cancer in well-fed people. The COSMOS trial (COcoa Supplement and Multivitamin Outcomes Study), a 21,442-person randomized trial in US adults aged 60 and over, has complicated that line. Its primary cancer and cardiovascular results were null, as expected, but its cognitive substudies and a 2026 epigenetic-clock analysis found small, consistent benefits from a cheap daily multivitamin-multimineral (MVM) tablet.

## What COSMOS found

Outcome| Result  
---|---  
Cancer and cardiovascular disease (primary)| No effect on invasive cancer (HR 0.97, 95% CI 0.86 to 1.09), cardiovascular events (HR 0.98) or all-cause mortality (HR 0.93, 95% CI 0.81 to 1.08) over a median 3.6 years; lung cancer was lower (HR 0.62), a secondary finding (PMID [35294969](https://pubmed.ncbi.nlm.nih.gov/35294969/))  
Cognition, three substudies pooled| Better global cognition (0.07 SD units, 95% CI 0.03 to 0.11) and episodic memory (0.06 SD units) across 5,203 participants in COSMOS-Clinic, COSMOS-Mind and COSMOS-Web; the authors equate the global effect to about 2 years less cognitive ageing (PMID [38244989](https://pubmed.ncbi.nlm.nih.gov/38244989/))  
Epigenetic clocks, 958 participants, 2 years| Slower yearly increase in PCGrimAge (-0.113 years, 95% CI -0.205 to -0.020) and PCPhenoAge (-0.214 years); the effect on PCGrimAge was concentrated in people who were ageing fast at baseline. First-generation clocks and DunedinPACE did not change, and cocoa extract had no effect on any clock (PMID [41803341](https://pubmed.ncbi.nlm.nih.gov/41803341/))  
  
## The contradicting evidence

The Physicians' Health Study II, which gave 5,947 male physicians aged 65 and over a daily multivitamin and tested cognition by telephone over up to 12 years, found **no cognitive benefit** : the difference in global cognitive change versus placebo was -0.01 SD units (95% CI -0.04 to 0.02). Its authors suggested the dose or the well-nourished population might explain it (PMID [24490265](https://pubmed.ncbi.nlm.nih.gov/24490265/)). The same trial found no reduction in cardiovascular events (PMID [23117775](https://pubmed.ncbi.nlm.nih.gov/23117775/)). So the cognitive question has one large trial for and one against, with COSMOS testing a broader population (women and men, not only physicians) using a more detailed test battery.

## How to read the effect sizes

The benefits are small. A 0.07 standard-deviation difference in cognition is real at the population level but would not be noticeable in an individual over two years. The clock effects amount to roughly six weeks of slower biological ageing per year on the clocks that moved. The most interesting detail is the subgroup: the PCGrimAge benefit appeared almost entirely in people whose biological age was already running ahead of their calendar age, consistent with the idea that a multivitamin mostly helps people with something to correct.

## Funding and interests

COSMOS investigators received investigator-initiated grants from Mars Edge (a Mars Inc. nutrition unit) for infrastructure and study pills, and Pfizer Consumer Healthcare (now Haleon) donated the multivitamin pills; the clock analysis also received in-kind methylation data from a clock company, and several authors report supplement-industry grants or honoraria. The cognitive substudies had NIH funding. All of this is disclosed in the papers (PMID [41803341](https://pubmed.ncbi.nlm.nih.gov/41803341/)). It does not invalidate the findings, but independent replication matters more than usual.

## Where it sits on the Heretic scale

**Promising** for slower cognitive and epigenetic ageing in adults over 60: randomized evidence, replicated across three cognitive substudies within one trial, small effects, and one conflicting large trial. Not supported: protection against cancer, heart disease or death. Related: [Omega-3 and biological aging](omega-3-and-biological-aging.html) (the DO-HEALTH clock result), [Biological age clocks](biological-age-clocks-measurement-and-interpretation.html) and [Supplements and nutraceuticals](supplements-and-nutraceuticals.html).

## Sources

1. Sesso HD, et al. (2022). "Multivitamins in the prevention of cancer and cardiovascular disease: the COcoa Supplement and Multivitamin Outcomes Study (COSMOS) randomized clinical trial." Am J Clin Nutr . PMID 35294969 . DOI 10.1093/ajcn/nqac056 . Trial NCT02422745 (PubMed 35294969, DOI 10.1093/ajcn/nqac056, NCT02422745)
2. Vyas CM, et al. (2024). "Effect of multivitamin-mineral supplementation versus placebo on cognitive function: results from the clinic subcohort of the COSMOS randomized clinical trial and meta-analysis of 3 cognitive studies within COSMOS." Am J Clin Nutr 119(3):692-701. PMID 38244989 . DOI 10.1016/j.ajcnut.2023.12.011 (PubMed 38244989, DOI 10.1016/j.ajcnut.2023.12.011)
3. Li S, et al. (2026). "Effects of daily multivitamin-multimineral and cocoa extract supplementation on epigenetic aging clocks in the COSMOS randomized clinical trial." Nat Med 32(3):1012-1022. PMID 41803341 . DOI 10.1038/s41591-026-04239-3 (PubMed 41803341, DOI 10.1038/s41591-026-04239-3)
4. Grodstein F, et al. (2013). "Long-term multivitamin supplementation and cognitive function in men: a randomized trial." Ann Intern Med . PMID 24490265 . DOI 10.7326/0003-4819-159-12-201312170-00006 (PubMed 24490265, DOI 10.7326/0003-4819-159-12-201312170-00006)
5. Sesso HD, et al. (2012). "Multivitamins in the prevention of cardiovascular disease in men: the Physicians' Health Study II randomized controlled trial." JAMA . PMID 23117775 . DOI 10.1001/jama.2012.14805 (PubMed 23117775, DOI 10.1001/jama.2012.14805)