Daily multivitamin nudged memory in older adults

Whether a cheap daily multivitamin-mineral supplement can slow late-life cognitive decline is one of the longest-running open questions in nutrition, and until recently the evidence was too thin for clinicians to make any recommendation. Earlier signals within large trials came mainly from telephone or web-based cognitive testing, which is convenient but less detailed than sitting a participant down for a full battery. What was missing was a test of longer-term multivitamin effects using in-person, detailed neuropsychological assessment, plus a pooled look across the separate cognitive studies run inside the same parent trial to see whether the signal holds up.

COSMOS is a 2 x 2 factorial randomised trial of cocoa extract providing 500 mg flavanols per day and/or a daily multivitamin-mineral supplement, for cardiovascular disease and cancer prevention, among 21,442 United States adults aged 60 years and over. The clinic subcohort, COSMOS-Clinic, comprised 573 participants who completed all cognitive tests administered in person at baseline and were followed for change over 2 years across global cognition, episodic memory, and executive function and attention. The authors then meta-analysed non-overlapping participants across three COSMOS cognitive substudies: COSMOS-Clinic (n = 573), COSMOS-Mind (n = 2158) and COSMOS-Web (n = 2472).

In COSMOS-Clinic the benefit on global cognition over 2 years was modest and not statistically clear, at 0.06 SD units (95% CI -0.003 to 0.13). The composite that did move was episodic memory, with a mean difference of 0.12 SD units (0.002 to 0.23). Executive function and attention showed 0.04 SD units (-0.04 to 0.11), which is a null result. The pooled analysis across the three substudies was more definite: global cognition improved by 0.07 SD units (0.03 to 0.11; P = 0.0009) and episodic memory by 0.06 SD units (0.03 to 0.10; P = 0.0007). The authors describe the global cognition effect as equivalent to reducing cognitive ageing by 2 years.

This strengthens the case built by the earlier COSMOS cognitive substudies by adding detailed in-person neuropsychological testing rather than remote assessment, and by pooling the substudies for a more precise estimate. The consistency of the episodic memory signal across different assessment methods is the most persuasive element. But all three substudies sit inside one parent trial in one country among adults aged 60 and over, so this is internal consistency rather than independent replication, and the effect sizes are small in absolute terms. Whether benefits persist beyond 2 years, translate into fewer dementia diagnoses, or generalise to younger or nutritionally different populations is unresolved.

This is a reason to think a daily multivitamin-mineral supplement produces a small, measurable advantage in memory-related cognitive scores in older adults over a couple of years, rather than nothing at all. The effect is modest, and in the in-person subcohort it did not extend to executive function or attention. Framing it as roughly two years less cognitive ageing is the authors' way of making a small standardised difference intelligible, not a promise about anyone's individual trajectory. Nothing here speaks to preventing dementia, to specific formulations or doses, or to adults much younger than the trial population.