Evidence File 033 · Multivitamins · Method v0.14 of 4 claims researched

MultivitaminsBroad-spectrum multivitamin-mineral · 4 claims

Settled on mortality and disease prevention. Genuinely unsettled on cognition, where two large randomised programmes disagree.

4Claims scored

6Sources

1 tabletStudied dose

Jul 26Reviewed

Claim 01

Reduces all-cause mortality

2.8/10

Evidence against·Tested repeatedly at very large scale

A 2024 analysis of multivitamin use and mortality across three prospective US cohorts covering over 390,000 people found no mortality benefit. A 2011 study of over 182,000 participants found no effect. COSMOS found no impact on death rates.

A rapid review of 19 meta-analyses concluded that multivitamin supplementation does not affect all-cause mortality, with no benefit across primary or secondary prevention populations.

The working

2Quantity

2Quality

1Consistency

9Directness

(2 × 0.25) + (2 × 0.35) + (1 × 0.25) + (9 × 0.15) = 2.8
Scored as support for the claim. Among the best-tested negative claims in nutrition.

RevMultivitamin and mineral use: a rapid review of meta-analyses on health outcomes 19 meta-analyses appraised · no effect on all-cause mortality

Claim 02

Prevents cardiovascular disease or cancer

3.9/10

Weak·COSMOS null · one modest cancer signal in PHS II

The main COSMOS trial, 21,442 participants over three years, found multivitamins did not protect against cardiovascular disease or cancer.

The Physicians’ Health Study II did find a modest 8% reduction in total cancer incidence among male physicians over 11 years — without any associated mortality benefit, and with the effect diminishing across tumour types. That finding has not led to population-wide recommendations.

The US Preventive Services Task Force concludes evidence is insufficient to recommend multivitamins for primary prevention.

The working

3Quantity

3Quality

3Consistency

9Directness

(3 × 0.25) + (3 × 0.35) + (3 × 0.25) + (9 × 0.15) = 3.9
Scored as support for the claim. A single modest positive on one endpoint against a large negative literature.

RevCOSMOS main trial outcomes summarised: no protection against cardiovascular disease or cancer 21,442 participants, over three years
RevMultivitamins in adult medical practice: evidence, risks, and pragmatic prescribing PHS II 8% cancer incidence reduction without mortality benefit; USPSTF position

Claim 03

Slows cognitive decline in older adults

5.6/10

Contested·COSMOS positive · PHS II null over 10 years

A meta-analysis of three cognitive substudies within COSMOS, covering more than 5,000 non-overlapping participants aged 60 and over, found significant improvements in global cognition (MD 0.07, 95% CI 0.03–0.11) and episodic memory (MD 0.06, 95% CI 0.03–0.10) — an effect the authors described as equivalent to reducing cognitive ageing by roughly two years.

The Physicians’ Health Study II found the opposite. In 5,947 male physicians followed an average of 10 years, there was no difference in mean cognitive change between multivitamin and placebo groups.

Two details matter for reading this. The COSMOS clinic subcohort on its own was not significant (MD 0.06, 95% CI −0.003 to 0.13) — the positive result comes from pooling. And the COSMOS authors propose the benefit appears in the first 2–3 years and might have been missed by PHS II’s longer follow-up. That is a reasonable hypothesis, not a settled reconciliation.

The working

6Quantity

6Quality

3Consistency

8Directness

(6 × 0.25) + (6 × 0.35) + (3 × 0.25) + (8 × 0.15) = 5.6
Consistency scores 3. Two large randomised programmes, opposite answers, and the disagreement has a plausible methodological explanation rather than a resolution.

MetaVyas CM, et al. Effect of multivitamin-mineral supplementation versus placebo on cognitive function: COSMOS-Clinic and meta-analysis of 3 COSMOS cognitive studies Am J Clin Nutr 2024 · doi 10.1016/j.ajcnut.2023.12.011 · n>5,000
RCTPhysicians’ Health Study II cognitive outcomes, as compared within the COSMOS report n=5,947, mean age 71.6, ~10 year follow-up, no MVM benefit observed

Claim 04

Fills genuine dietary gaps in specific groups

5.8/10

Emerging·Benefit concentrated where baseline status is poor

The rapid review of 19 meta-analyses found benefits concentrated in specific groups: reduced psychological symptoms in stressed younger adults, and reduced episodes of infection in younger, undernourished or hospitalised adults.

The Physicians’ Health Study II similarly suggested baseline nutritional status modifies response, with greater benefit in those with lower dietary quality. The consistent theme across the whole multivitamin literature is that supplementation helps where something is actually missing, and does very little where nothing is.

The working

5Quantity

5Quality

7Consistency

7Directness

(5 × 0.25) + (5 × 0.35) + (7 × 0.25) + (7 × 0.15) = 5.8
Consistent across reviews in direction, and the populations concerned are narrow and not the general supplement market.

RevMultivitamin and mineral use: rapid review of meta-analyses, population subgroup findings Benefits in stressed younger adults and undernourished or hospitalised adults

Dosage as studied

Most-studied regimen

1 tablet daily, standard formulation

COSMOS used Centrum Silver supplied by the manufacturer. Trial doses are standard retail formulations, not megadoses.

COSMOS: 21,442 adults, women 65+ and men 60+, randomised double-blind placebo-controlled 2×2 factorial, over three years. MVM supplied by Haleon.
Physicians’ Health Study II: 5,947 male physicians, mean age 71.6, average 10-year follow-up, running 1997 to 2011.
USPSTF position: Evidence insufficient to recommend multivitamins for primary prevention; advises against beta carotene or vitamin E for cardiovascular prevention.
Where benefit does appear: Specific populations — stressed younger adults, and undernourished or hospitalised adults — rather than the general well-nourished public.