Multivitamins and memory: how much did COSMOS change?

A modest result can be useful without becoming an anti-aging claim

COSMOS gives multivitamins a more interesting cognitive evidence base than a generic wellness slogan. Several randomized cognitive studies within the project found small advantages in test performance with a daily multivitamin-mineral supplement. That is worth examining carefully. It does not establish that a bottle of vitamins prevents dementia, restores lost independence or makes a person biologically younger. The distinction matters when a modest result becomes the headline for a premium product. ^{[1] [2] [3]}

The sensible buying question is what the tested intervention and observed outcome establish together. A familiar ingredient list is not enough to transfer a result to a different formulation. A statistically significant group difference also need not be noticeable to every participant in everyday life. Start with the test, the size of the difference and the follow-up period before deciding how much importance to place on a supplement in your own routine.

Another intervention with a specific cognitive trial population is covered in hearing intervention and cognition in ACHIEVE.

COSMOS-Web tested recall with an online battery

COSMOS-Web included 3,562 older adults assigned a daily Centrum Silver multivitamin or placebo and assessed with an internet-based cognitive battery over three years. Its prespecified primary outcome was the one-year change in immediate word recall. The multivitamin group performed better on that outcome and on the average recall comparison across three years. The report did not find significant effects on the other secondary cognitive outcomes. ^[1]

The authors translated the recall difference into an estimated 3.1 years of age-related memory change using a cross-sectional age-performance relationship. ^[1] That is a comparison on a particular test scale, not three years of added life or evidence that aging ran backward. It does not mean a participant gained three years of memory in every context. Treat age-equivalent descriptions as explanatory models; keep the actual measured task visible whenever a product cites the study.

Telephone assessment provided another small signal

COSMOS-Mind randomized 2,262 participants and used telephone cognitive assessments over three years. The study tested both cocoa extract and multivitamin-mineral supplementation in a factorial design. Cocoa extract did not significantly improve global cognition. The multivitamin comparison favored supplementation by 0.07 standardized units, with a 95% confidence interval of 0.02 to 0.12. The investigators also reported benefits in episodic memory and executive function. ^[2]

The size of the effect deserves as much attention as its statistical significance. A standardized unit expresses the difference relative to variation in the measured scores; it is not a percentage improvement in intelligence or a probability that an individual avoids dementia. The finding supports a narrow statement about average cognitive performance under the study conditions. It does not identify which nutrient produced the difference or show that adding more of a particular ingredient would make the effect larger.

The clinic study and pooled analysis need separate readings

In COSMOS-Clinic, 573 participants completed the baseline cognitive battery. Over two years the global-cognition difference was 0.06 standardized units, with a confidence interval from -0.003 to 0.13, crossing zero. Episodic memory favored supplementation by 0.12 units, confidence interval 0.002 to 0.23; executive function and attention did not show a significant difference. These details are more informative than calling every cognitive domain positive. ^[3]

The accompanying meta-analysis used nonoverlapping participants across the three cognitive substudies. Its global-cognition difference was 0.07 standardized units, confidence interval 0.03 to 0.11; episodic memory differed by 0.06 units, interval 0.03 to 0.10. ^[3] Pooling avoids counting overlapping participants twice, but these remain studies within the same larger project. They should not be described as three wholly unrelated replications across every kind of older adult, supplement and assessment setting.

Dementia prevention remains a different outcome

A separate COSMOS-Mind analysis adjudicated 110 incident cases of mild cognitive impairment and 14 cases of probable dementia over three years. Incidence did not differ significantly by multivitamin or cocoa-extract assignment. The authors cautioned that statistical power was low. ^[4] This does not prove that a benefit is impossible; it means these data did not demonstrate reduced incidence. A small number of dementia cases cannot support a confident preventive claim.

For a reader, the distinction is concrete. Doing somewhat better on a memory task and avoiding a diagnosis that affects daily function are both meaningful questions, but one result cannot substitute for the other. Persistent memory concerns still need appropriate assessment. Buying a supplement should not become a reason to postpone that conversation or reinterpret new difficulties as something a product subscription will automatically handle.

The product, total intake and practical purpose still matter

The NIH Office of Dietary Supplements advises checking total intake from fortified foods and other supplements so that adding a multivitamin does not push nutrients above their upper limits. It also highlights vitamin K interactions with warfarin and similar anticoagulants. ^[5] Bring the actual ingredient panel and the other products you use to a medication or nutrition review. “Basic multivitamin” can conceal different amounts and combinations, especially when several products overlap.

My reading of COSMOS is favorable toward the specificity of the question and cautious about the marketing that can follow. The evidence supports discussing a modest average cognitive-test benefit from the studied regimen in older adults. It does not rank greens powders, personalized packs or high-dose longevity stacks. Choose a supplement for a defined nutritional or practical reason, at a price you can justify, while keeping its evidence limits clear. The more expensive product should have to establish its own value rather than borrowing an unrelated formulation’s trial.

For the difference between a basic multivitamin and a broader greens product, read our AG1 review.

For another example of a nutrient being tested against an actual health outcome, see vitamin D and fracture prevention.