Multivitamin Supplementation and Cognitive Decline: What the COSMOS Studies Found

Cognitive decline is one of the most significant concerns associated with ageing — the gradual erosion of memory, processing speed, and executive function that can begin decades before any clinical diagnosis. Research into this area has accelerated substantially, and one of the most studied interventions is a supplement that millions of people already take every day: a multivitamin and multimineral.

What makes this topic genuinely compelling is the quality of the evidence behind it. Multivitamins have been the subject of large-scale, rigorous randomised trials — and the findings on cognitive function are more nuanced, and more interesting, than the headlines often suggest. The evidence does not support using multivitamins to reduce cancer risk or prevent cardiovascular disease. But for cognitive outcomes specifically, the picture is more encouraging — and the distinction between these two questions is often obscured in popular coverage of the research.

This article examines the COSMOS programme trials in detail: what they measured, how large the effect was, why an earlier negative study produced a different result, and what a 2024 viral study actually showed (and did not show). It also covers the evidence on which ingredients to seek — and which to avoid — when selecting a multivitamin formulation.

The Source of the Claims: the COSMOS Studies

The first question any sceptical reader should ask is: what is the evidence that a multivitamin improves brain performance at all? Extraordinary claims require extraordinary evidence.

In this case, the evidence comes from the COSMOS programme — a large-scale research initiative that generated three ancillary studies: COSMOS-Mind, COSMOS-Web, and COSMOS-Clinic. Data from all three studies were combined into a meta-analysis involving over 5,000 people.

All three studies were randomised, placebo-controlled, and double-blind. This means that participants were randomly assigned to receive either the multivitamin/multimineral supplement or an identical-looking placebo, and neither the participants nor the researchers knew which was which. This design is the gold standard for minimising bias and producing reliable results.

Two of the three studies (COSMOS-Mind and COSMOS-Clinic) examined Global Cognition and together enrolled 2,731 participants. The results reached a P value of 0.0009 — a level of statistical significance indicating the finding is extremely unlikely to be due to chance.

The third study, COSMOS-Web, focused on episodic memory and likewise found significant results. When all three studies' data were combined in the meta-analysis — encompassing over 5,000 participants — the pooled P value was 0.0007, strongly indicating a genuine benefit from the supplement.

Another critically important aspect of these results: the effect was not heterogeneous across the studies. Each study individually showed the same directional finding, which means the results are reproducible — a key test of scientific reliability. Heterogeneity across studies (where some show an effect and others do not) is often a sign that the finding is not robust. The consistency here strengthens the interpretation that the result reflects a genuine biological effect rather than statistical noise.

The COSMOS programme is notable for its scale and design rigour. Large randomised controlled trials of this nature are expensive and difficult to run, which is why the convergence of results across three independent but coordinated studies within the same programme adds meaningful confidence to the findings.

Measuring the Impact

Demonstrating that an effect exists is important, but so is understanding the size of that effect. How meaningful, in practical terms, is the cognitive benefit from multivitamin supplementation?

Cognitive function is inherently complex to quantify, but the COSMOS researchers framed the benefit in a clinically accessible way: multivitamin supplementation was associated with an effect equivalent to approximately two years less cognitive ageing compared to the placebo group.

"Regarding the clinical importance of our meta-analysis findings: the magnitude of the effect of MVM supplementation compared with placebo on global cognition score was equivalent on average to the difference observed between participants 2 y apart in age (that is, 2 y less cognitive aging)." — COSMOS

Given that the trials ran for only 2–3 years, an effect of this magnitude is substantial. It is also where the first controversies enter the picture — because not all the evidence points in the same direction.

The First Controversy

The supplement at the centre of this debate is a multivitamin and multimineral — not a novel or exotic compound, but one of the most widely used dietary supplements in the world.

The controversy arises in part because the COSMOS findings contradict an earlier, rigorous long-term study. The Physicians' Health Study II was also a randomised, double-blind, placebo-controlled trial, and it enrolled nearly 6,000 male physicians followed from 1997 to 2011 — a substantially longer follow-up than COSMOS.

In contrast to the COSMOS studies, the Physicians' Trial found "no difference in mean cognitive change over time between the multivitamin and placebo groups or in the mean level of cognition."

So, two long-term, double-blind, placebo-controlled trials reached opposite conclusions. What explains the discrepancy?

There are three major methodological differences between the studies, and the third is arguably the most important.

The first difference is participant composition: the Physicians' study enrolled only male participants, while the COSMOS studies enrolled both men and women — a broader and more representative sample.

The second difference is timing: the Physicians' study did not begin cognitive testing until, on average, two and a half years after supplementation started. Any cognitive benefits that accrued during the first 2.5 years would therefore have been missed entirely by the primary outcome measurement.

The third and most consequential difference is the supplement itself. The Physicians' study used Centrum Silver. The COSMOS trials used a different formulation — one that included two nutrients absent from Centrum Silver: lutein and lycopene. These carotenoids have emerging evidence for supporting brain health, and their presence in the COSMOS formula may partly account for the different outcomes.

The Viral Study

A separate controversy emerged from a 2024 study that attracted significant media coverage: the Multivitamin Use and Mortality Risk in 3 Prospective US Cohorts analysis, involving over 390,000 participants.

This study adds to a body of prior research on the same question. A 2011 study involving over 182,000 participants found no effect of multivitamin use on mortality. A 2022 analysis reached the same conclusion — insufficient evidence for multivitamins in preventing cancer or cardiovascular disease. The COSMOS trials themselves also found no impact on cancer, cardiovascular disease, or all-cause mortality. The 2024 study confirmed this pattern once again: no benefit from multivitamins on cancer, cardiovascular disease, or general death rates.

Some news outlets and medical commentators cited the 2024 study as evidence that multivitamins are "useless" across the board. This interpretation misrepresents what the study actually measured. The 2024 study — and all the mortality/cardiovascular studies before it — did not examine cognitive outcomes. The COSMOS studies, which specifically tested cognitive function, reached a different conclusion.

The research consistently shows that multivitamin supplementation does not meaningfully reduce the risk of cancer, heart disease, or premature death in people with an otherwise adequate diet. That is the settled scientific position, and the COSMOS studies do not contradict it. What the COSMOS data suggest — separately and distinctly — is a measurable benefit for cognitive function over 2–3 years of supplementation.

These are different outcomes, measured in different studies, and conflating them leads to poor health decisions in either direction. Those seeking a multivitamin to reduce cancer risk or extend lifespan will find the evidence disappointing — the data consistently does not support that use. Those interested in supporting long-term cognitive function have a more substantive evidence base to consider. Understanding what the research actually measured — and what it did not — is essential to interpreting it correctly.

How to Pick a Multivitamin and Multimineral Supplement

The evidence suggests that not all multivitamins are equivalent — formulation matters, as the COSMOS-vs-Physicians comparison illustrates. Here is what the research suggests to look for, and what to avoid.

A fundamental principle is appropriate dosing. The body requires vitamins and minerals within defined ranges. Below those ranges, deficiency symptoms emerge; above them, some nutrients carry risks of toxicity or adverse effects. A well-designed supplement should fill genuine dietary gaps without megadosing — that is, without providing far more than the recommended daily intake of any nutrient.

Megadosing is sometimes a clinical tool (e.g., high-dose iron for documented iron-deficiency anaemia), but for a general daily supplement taken by people with reasonably balanced diets, it is rarely justified and can be counterproductive. A supplement should supplement a healthy diet, not replace it or override it with extreme nutrient loads.

The COSMOS study used an appropriately dosed formulation — which is one reason its results are more likely to be generalisable than a megadose product.

Two specific ingredients are worth avoiding in a general-purpose daily multivitamin, based on current evidence:

Beta-carotene (a precursor to Vitamin A): The US Preventive Services Task Force 2022 recommendations concluded that supplemental beta-carotene supplementation carries enough risk to outweigh its benefits for most people. Most diets that include vegetables already provide adequate Vitamin A, so a supplement containing beta-carotene is likely to push total intake above the optimal range.

Vitamin E: The same USPSTF guidance recommended against supplemental Vitamin E. A Cochrane review found that higher doses of both Vitamin E and Vitamin A were associated with increased mortality risk. For individuals with a reasonably varied diet, adding Vitamin E via a supplement is likely unnecessary and potentially harmful.

On the positive side, the COSMOS formula and the mechanistic literature point toward several ingredients that may be worth including in a multivitamin for cognitive support and overall health:

• Vitamin K2 — distinct from K1, and relevant to cardiovascular and bone health
• Magnesium Taurate — a chelated form of magnesium with good bioavailability; see the detailed write-up on common magnesium supplementation mistakes
• Trimethylglycine (TMG) — a methyl donor with emerging research on homocysteine metabolism
• Hyaluronic Acid — studied for joint and skin support
• Lutein and lycopene — the carotenoids that distinguished the COSMOS formula from the Centrum Silver formula used in the Physicians' study; among the most plausible candidates for explaining the cognitive benefit differential

When evaluating any multivitamin, it is worth checking the full ingredient list against this framework: appropriate doses, no beta-carotene, no high-dose Vitamin E, and the presence of well-evidenced ingredients relevant to your personal health goals. The formulation that produced the COSMOS results is a useful benchmark.

The lutein and lycopene question is particularly relevant for anyone choosing a multivitamin based on the COSMOS findings. If the COSMOS formula's inclusion of these carotenoids contributed to its cognitive benefit — and that is a plausible mechanistic hypothesis — then a multivitamin that omits them may not replicate the COSMOS result, regardless of how well-formulated it is in other respects. The evidence base for lutein's role in cognitive support, particularly via macular pigment density and central nervous system carotenoid concentrations, has strengthened in recent years, making its inclusion in cognitive-health-focused formulations increasingly justifiable.

As with any supplement decision, individual health circumstances vary. Reviewing the full ingredient list and discussing supplementation with a qualified healthcare professional is the appropriate path before starting any new regimen. The research summarised here provides a framework for that conversation — the key questions being: does the formulation avoid beta-carotene and high-dose Vitamin E, does it include the evidence-supported nutrients at appropriate doses, and does it align with the COSMOS formula design rather than departing from it?

Sources:

1. Effect of multivitamin-mineral supplementation versus placebo on cognitive function: results from the clinic subcohort of the COcoa Supplement and Multivitamin Outcomes Study (COSMOS) randomized clinical trial and meta-analysis of 3 cognitive studies within COSMOS: https://www.sciencedirect.com/science/article/abs/pii/S0002916523663427?via%3Dihub
2. Long-Term Multivitamin Supplementation and Cognitive Function in Men: A Randomized Trial: https://www.acpjournals.org/doi/10.7326/0003-4819-159-12-201312170-00006
3. Multivitamin Use and the Risk of Mortality and Cancer Incidence: The Multiethnic Cohort Study: https://academic.oup.com/aje/article/173/8/906/156839
4. Vitamin, Mineral, and Multivitamin Supplementation to Prevent Cardiovascular Disease and Cancer, US Preventive Services Task Force Recommendation Statement: https://jamanetwork.com/journals/jama/fullarticle/2793446
5. Multivitamins in the prevention of cancer and cardiovascular disease: the COcoa Supplement and Multivitamin Outcomes Study (COSMOS) randomized clinical trial: https://pubmed.ncbi.nlm.nih.gov/35294969/
6. Multivitamin Use and Mortality Risk in 3 Prospective US Cohorts: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11208972/
7. Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases: https://www.cochrane.org/CD007176/LIVER_antioxidant-supplements-for-prevention-of-mortality-in-healthy-participants-and-patients-with-various-diseases