Multivitamin / Multivitamin-Mineral (MVM) for Cognitive Function
A daily multivitamin may modestly slow age-related cognitive decline in older adults, but the effect is very small, comes almost entirely from one trial family using a single product, and is contradicted by other long-term data, so it is not a reliable way to protect memory or prevent dementia.
Why this grade7-layer evidence engine
The grade rests on conflicting high-quality evidence. The three COSMOS substudies, all using daily Centrum Silver, pointed in the same direction: COSMOS-Mind (PMID 36102337, n=2,262) improved global cognition by 0.07 SD (p=0.007), COSMOS-Web (PMID 37236815, n=3,562) improved memory recall (p=0.004), and a 2024 meta-analysis of all three (PMID 38244989, n=5,200) pooled to SMD 0.07 (p=0.0009), roughly two years of slower cognitive aging.
That signal is weak and disputed rather than convincing. The effect size (Cohen's d about 0.07) is far below the usual threshold for a small effect, every positive trial used the same product and shared industry funding, and the larger, longer Physicians' Health Study II (PMID 24490265, n=5,947, ~12 years) found nothing (mean difference -0.01, p=0.54). An older meta-analysis (PMID 22330823) saw a benefit only on immediate recall, and PREADViSE (PMID 28944218) was null for vitamin E and selenium. No trial has shown reduced dementia incidence.
Authorities remain cautious or unconvinced, which holds the rating at C. The FDA grants multivitamins no cognitive health claim, and the NHS advises that a varied, balanced diet supplies what most people need; WHO endorses micronutrient supplements only in antenatal care, not for cognition. Cleveland Clinic states plainly that for men 65 and older multivitamins provide no cognitive benefit. Any benefit is best read as limited to the Centrum Silver evidence in older adults, not a general memory or anti-dementia aid.
Scoring transparency
All scores computed by a 7-layer evidence engine — fully auditable▸View the full decision path (audit trail)
- compute_raw_score — 加權公式: L2×0.30 + L3×0.25 + L5×0.25 + L11×0.10 + L1×0.10 = 0.603
- tier_from_score — 依分數區間映射至 tier letter
- apply_hec_rules — 高品質 SR/MA 顯示 positive (1 篇 > 0 negative)
- tier_strict_requirement_check — | B→C 因 scope.conflation_risk=true 且 L11 獨評較低 (B7-2 tier cap)
- detect_disputes — 偵測到 1 個 hard + 0 個 soft dispute
- decide_status — 依 tier + dispute 結果決定 status