INDEPENDENT THINKING. WHOLE-FOOD LIVING.NO ANIMALS / NO OIL
THENANODIET
← Nutrition & Adequacy / Health Desk

Cocoa extract and cognition: no benefit in a three-year randomized trial

COSMOS-Mind found no global cognitive benefit from cocoa extract, distinct from its multivitamin comparison.

Whole-food ingredients prepared without added oil

Illustrative food image, not a photograph of the study intervention.

A three-year randomized question

COSMOS-Mind tested cocoa extract rather than a cocoa-containing meal. Its randomized two-by-two factorial design compared daily cocoa extract with placebo and, separately, a commercial multivitamin-mineral with placebo. The cocoa preparation contained 500 mg of flavanols per day, and the study lasted three years. Cognition was assessed by telephone at baseline and annually. These details define the intervention: this was not an experiment assigning people a chocolate dessert or an ordinary cocoa beverage. The amount describes the research preparation, not a recommended intake. The primary endpoint was change in a global cognition composite with three years of cocoa extract use.

Baker LD, Manson JE, Rapp SR, Sesso HD et al.. Alzheimer's & dementia : the journal of the Alzheimer's Association, 2023. PMID 36102337 · DOI

The cocoa result was not significant

The trial enrolled 2,262 participants with a mean age of 73 years. For cocoa extract, the reported global cognition effect was a mean z-score of 0.03, with a 95% confidence interval from -0.02 to 0.08 and P = .28. The abstract concludes that cocoa extract did not benefit cognition. The estimate and its uncertainty belong together: the interval includes zero, and the finding does not support a claim of improved global cognition from this preparation. This standardized test result is not a percentage change in memory, nor is it a measured reduction in dementia incidence.

Baker LD, Manson JE, Rapp SR, Sesso HD et al.. Alzheimer's & dementia : the journal of the Alzheimer's Association, 2023. PMID 36102337 · DOI

Do not transfer the other intervention’s result

The multivitamin-mineral comparison was a prespecified secondary endpoint, separate from the primary cocoa question. Relative to placebo, its global cognition effect was a mean z-score of 0.07, with a 95% confidence interval from 0.02 to 0.12 and P = .007. The abstract also reports multivitamin-mineral benefits for memory and executive function. Those results cannot be attributed to cocoa: the cocoa-by-multivitamin interaction was not significant for any cognitive composite. The authors call for confirmation in a more diverse cohort and further work on mechanisms. This article therefore preserves the distinction between two interventions rather than turning one positive comparison into a cocoa benefit.

Baker LD, Manson JE, Rapp SR, Sesso HD et al.. Alzheimer's & dementia : the journal of the Alzheimer's Association, 2023. PMID 36102337 · DOI

Keep the population and endpoint attached

Women made up 60% of participants, and 89% were non-Hispanic White. The abstract reports that 92% completed baseline and at least one annual assessment; that is not a statement that 92% completed every assessment over three years. These characteristics matter when considering how broadly to apply the findings, particularly given the authors’ request for confirmation in a more diverse cohort. The primary composite combined standardized scores from individual cognitive tests, and the analysis used intention-to-treat. Its outcome was cognitive test performance. It does not establish prevention of dementia, and a cocoa-containing recipe should not be presented as the extract tested here.

Baker LD, Manson JE, Rapp SR, Sesso HD et al.. Alzheimer's & dementia : the journal of the Alzheimer's Association, 2023. PMID 36102337 · DOI