Unexpected results from a phase 3 trial exploring the effect of multivitamins and cognition have now been published.
Findings from a phase 3 study show daily multivitamin use, but not cocoa, is linked to a significantly slower rate of age-related cognitive decline.
Originally presented last November at the 14th Clinical Trials on Alzheimer’s Disease (CTAD) conference and reported by Medscape Medical News at that time, this is the first large-scale, long-term randomized controlled trial to examine the effects of cocoa extract and multivitamins on global cognition. The trial’s primary focus was on cocoa extract, which earlier studies suggest may preserve cognitive function. Analyzing the effect of multivitamins was a secondary outcome.
Showing vitamins, but not cocoa, was beneficial is the exact opposite of what researchers expected. Still, the results offer an interesting new direction for future study, lead investigator Laura D. Baker, PhD, professor of gerontology and geriatric medicine at Wake Forest School of Medicine, Winston-Salem, North Carolina, told Medscape Medical News.
“This study made us take notice of a pathway for possible cognitive protection,” Baker said. “Without this study, we would never have looked down that road.”
Unexpected Effect
The COSMOS-Mind study is a sub-study to a larger parent trial called COSMOS. It investigated the effects of cocoa extract and a standard multivitamin-mineral (MVM) on cardiovascular and cancer outcomes in more than 21,000 older participants.
In COSMOS-Mind, researchers tested whether daily intake of cocoa extract vs placebo and a multivitamin-mineral vs placebo improved cognition in older adults.
More than 2200 participants aged 65 and older were enrolled and followed for 3 years. They completed tests over the telephone at baseline and annually to evaluate memory and other cognitive abilities.
Results showed cocoa extract had no effect on global cognition compared with placebo (mean z-score, 0.03; P = .28). Daily multivitamin use, however, did show significant benefits on global cognition vs placebo (mean z, 0.07, P = .007).