Effect of multivitamin-mineral supplementation versus placebo on cognitive function: results from the clinic subcohort of the COcoa Supplement and Multivitamin Outcomes Study (COSMOS) randomized clinical trial and meta-analysis of 3 cognitive studies within COSMOS

Abstract

Background: Longer effects of multivitamin-mineral (MVM) supplementation on late-life cognitive function remain untested using in-person, detailed neuropsychological assessments. Furthermore, insufficient evidence exists for healthcare providers to recommend daily MVM supplements to prevent cognitive decline.

Objectives: This study aimed to test MVM effects on cognitive change using in-person, detailed neuropsychological assessments and conduct a meta-analysis within COSMOS (COcoa Supplement and Multivitamin Outcomes Study) cognitive substudies for a robust evaluation of MVM effects on cognition.

Methods: COSMOS is a 2 × 2 factorial trial of cocoa extract (500 mg flavanols/d) and/or a daily MVM supplement for cardiovascular disease and cancer prevention among 21,442 United States adults aged ≥60 y. There were 573 participants in the clinic subcohort of COSMOS (that is, COSMOS-Clinic) who completed all cognitive tests administered at baseline. For the meta-analysis, we included nonoverlapping participants across 3 COSMOS cognitive substudies: COSMOS-Clinic (n = 573); COSMOS-Mind (n = 2158); COSMOS-Web (n = 2472).

Results: In COSMOS-Clinic, we observed a modest benefit of MVM compared with placebo on global cognition over 2 y {mean difference [95% confidence interval (CI)] = 0.06 SD units (SU) (-0.003, 0.13)}, with a significantly more favorable change in episodic memory [mean difference (95% CI) = 0.12 SU (0.002, 0.23)] but not in executive function or attention [mean difference (95% CI) = 0.04 SU (-0.04, 0.11)]. The meta-analysis of COSMOS substudies showed clear evidence of MVM benefits on global cognition [mean difference (95% CI) = 0.07 SU (0.03, 0.11); P = 0.0009] and episodic memory [mean difference (95% CI) = 0.06 SU (0.03, 0.10); P = 0.0007]; the magnitude of effect on global cognition was equivalent to reducing cognitive aging by 2 y.

Conclusions: In COSMOS-Clinic, daily MVM supplementation leads to a significantly more favorable 2-y change in episodic memory. The meta-analysis within COSMOS cognitive substudies indicates that daily MVM significantly benefits both global cognition and episodic memory. These findings within the COSMOS trial support the benefits of a daily MVM in preventing cognitive decline among older adults. This trial was registered at COSMOS-clinicaltrials.gov as NCT02422745, at COSMOS-Mind-clinicaltrials.gov as NCT03035201, and at COSMOS-Web-clinicaltrials.gov as NCT04582617.

Keywords: cognition; geriatric; meta-analysis; multivitamin-multimineral supplementation; randomized clinical trial.

FIGURE 1

Flow of participants in the COSMOS-Clinic cognitive substudy. CE, cocoa extract; COSMOS, the COcoa Supplement and Multivitamin Outcomes Study [[14], [15], [16]]; MVM, multivitamin-mineral.

FIGURE 2

Adjusted mean difference in change in global cognition composite over a 2-y follow-up comparing MVM and placebo groups, according to prespecified subgroups.¹ 3MS, Modified Mini-Mental State [19]; AHEI, alternative healthy dietary index [29]; CE, cocoa extract; CI, confidence interval; CVD, cardiovascular disease; MVM, multivitamin-mineral; SU, SD units. ¹Analyses were performed using repeated measures models adjusted for baseline age, sex, randomization to the CE group, and CE effect over time (that is, CE-x-time). All participants contributed to the repeated measure analysis at one and/or both time points. Adjusted mean differences (95% CI) between MVM and placebo groups in global cognition scores over a 2-y follow-up period are shown within subgroups. Analyses were not adjusted for multiple comparisons. ²P-interaction is from the 1-df test of the subgroup-x-treatment-x-follow-up time interaction term in the model. ³In the COSMOS prerandomization questionnaire, participants were presented with the following categories of race and asked to check all that apply: American Indian or Alaska Native; Black or African American; Asian; Native Hawaiian or other Pacific Islander; White; Unknown. Regarding ethnicity, participants were asked to select one of these options: Hispanic; non-Hispanic; Unknown. On the basis of the responses, we derived a binary variable for self-reported racial and ethnic background, with categories of non-Hispanic White and other racial and ethnic groups [Black or African American; Hispanic (not Black or African American); Asian; Native Hawaiian or other Pacific Islander; American Indian or Alaska Native; multiple race or unknown race and/or unknown ethnicity]. ⁴On the basis of the self-report of transient ischemic attack, congestive heart failure, coronary artery bypass graft, angioplasty, or stent. ⁵For this subgroup analysis, we excluded the z-score of the general cognition test from the global cognition composite. ⁶Subjective cognitive complaints were ascertained using the Structured Telephone Interview for Dementia Assessment (STIDA) [30].

FIGURE 3

Meta-analyses of findings of global cognition and episodic memory, comparing MVM and placebo groups, among nonoverlapping participants across COSMOS cognitive substudies.¹ CI, confidence interval; COSMOS, COcoa Supplement and Multivitamin Outcomes Study [[14], [15], [16]]; MVM, multivitamin-mineral; SU, SD units. ¹Meta-analysis was performed using random-effects models. A pooled summary effect estimate with 95% (CI) was reported. Heterogeneity across substudies (using the Q statistic) and between-study variance [using tau squared (τ²)] were examined using the %METAANAL macro (DerSimonian–Laird method) [31,32]. P < 0.025 was used to indicate statistical significance for each coprimary outcome after Bonferroni correction [33].