In the COSMOS multivitamin trials, which subgroups (nutritional status, cardiovascular disease, race/ethnicity) showed the largest memory benefit?
This fact-check may be outdated. Consider refreshing it to get the most current information.
Executive summary
The COSMOS cognitive substudies found that daily multivitamin‑mineral (MVM) supplementation produced small but statistically significant improvements in episodic memory and global cognition overall, and pre‑specified subgroup analyses consistently pointed to the largest relative benefit among participants with a baseline history of cardiovascular disease (CVD) [1] [2]. The trial reports and meta‑analysis do not provide robust, published evidence that baseline nutritional deficiency or race/ethnicity subgroups experienced larger memory gains, and available publications either do not report those comparisons or found benefits broadly across participants [3] [4].
1. Overall memory benefit in COSMOS and how big it was
Across COSMOS substudies, daily MVM use for up to three years yielded modest improvements in episodic memory and global cognition: the pooled meta‑analysis of the clinic, Mind, and Web subcohorts showed a mean difference of about 0.06–0.07 standard‑unit (SU) improvements for episodic memory and global cognition, an effect the authors equated to roughly two years less cognitive aging for global cognition [1] [5]. COSMOS‑Web independently found improved verbal memory at one year that was sustained over three years in the ancillary web cohort of >3,500 participants [6] [7].
2. Cardiovascular disease history: the subgroup with the clearest larger memory gain
Pre‑specified subgroup analyses in COSMOS‑Mind and related reports repeatedly flagged participants with a history of CVD as showing a more pronounced MVM benefit, suggesting either greater relative improvement or protection from CVD‑related cognitive decline; investigators highlighted this pattern in the main Mind paper and follow‑up analyses [2] [3] [1]. Several summaries and press statements emphasize that the MVM benefit on cognitive composites — particularly episodic memory — was most marked among those with baseline CVD, and that participants with prior CVD on placebo tended to show continued decline whereas those on MVM showed less decline [8] [9].
3. Nutritional status: an unanswered or underreported question
Investigators explicitly stated that understanding whether particular subgroups defined by nutritional status benefit more or less is important, but the publicly available COSMOS cognitive papers and summaries included here do not present clear, published subgroup results demonstrating that baseline micronutrient deficiency or lower nutritional status modified the memory benefit [7] [4]. The trial materials note that subgroup analyses were selected a priori for factors important to cognitive decline, but the sources provided either omit detailed nutritional‑status interactions or focus reporting on CVD and overall effects, leaving the nutritional‑status question unresolved in the cited literature [4] [3].
4. Race/ethnicity: no clear evidence of differential memory effects in the provided reports
The accessed COSMOS cognitive publications and press materials reviewed here do not report that race or ethnicity subgroups showed the largest memory benefit, nor do they provide detailed race/ethnicity‑stratified effect estimates in the excerpts available; reporting centers on overall results and the CVD subgroup instead [5] [6] [4]. Because race/ethnicity analyses are not described in the supplied snippets, it is not possible from these sources to assert either the presence or absence of larger benefits in any racial or ethnic subgroup; that remains an information gap in the documents provided [4] [3].
5. Alternative interpretations, limitations, and implicit agendas to note
Although COSMOS authors and institutional press releases highlight the CVD‑related signal and overall memory gains, the magnitude of the effect is modest (fractions of a standard unit) and the studies are ancillary subcohorts of a large pragmatic trial designed primarily for cardiovascular and cancer endpoints, so interpretation as a clinical recommendation requires caution [1] [5]. Industry‑friendly outlets and trade groups have amplified the findings with strong takeaways emphasizing multivitamin benefit, which reflects an implicit agenda to promote supplements, while the investigators themselves call for mechanistic work and replication and note that longer‑term clinical dementia outcomes remain unproven [9] [2] [10]. Finally, the available reports do not resolve whether baseline nutritional deficiency or demographic variables like race/ethnicity materially modify the memory benefit, so those remain open questions for future COSMOS analyses or independent trials [7] [4].