multivitamins may improve cognition

 A recent large study found that a commercial multivitamin-mineral (MVM, Centrum Silver) tablets improved cognitive function (see cognitive function inc with MVI AlzDem2022 in dropbox, or DOI: 10.1002/alz.12767 ). thanks to Karen Henley for bringing this to my attention

 

Details:

-- the COSMOS-Mind study (Cocoa Supplement and Multivitamin Outcomes Study of the Mind, COSMOS-m), an ancillary study to a large pragmatic, placebo-controlled, 2x2 factorial clinical trial of cocoa extract (CE) and/or MVM on cardiovascular and cancer outcomes. patients were enrolled from 2016-17 [a pragmatic trial is more of a real-world trial, often done by phone interviews of patients, instead of a more structured efficacy trial]

    -- CE contained 500mg of cocoa flavanols including 80 mg (-)-epicatechins, 50 mg theobromine, and 15 mg caffeine

-- 2262 participants (mostly from the Brigham and Women's Hospital in Boston and the Women’s Health Initiative): mean age 73, 60% women, 89% non-Hispanic white, BMI 28

-- hyptertension 59%, CVD history 9%, depression 21%, NSAID use 28%, statin 44%

-- smoking: never 52%/past 45%/current 3%; alcohol intake: rare 29%/monthly 7%/weekly 37%/daily 27%; chocolate intake: rare 17%/monthly 14%/weekly 58%/daily 11%

-- baseline multivitamin use 43%

-- exercise: total Metabolic Equivalent (MET) hours of exercise/wk: 22 hours

-- exclusion criteria included: no history of MI, stroke, or cancer; inclusion criteria included >65yo and diabetes but not on insulin

-- tests done: 50-point Telephone Interview of Cognitive Status with 10-minute short delay word list recall, an additional 40-minute Long Delay Word List Recall, immediate and delayed Story Recall (with 55% more components to remember than prior versions of the test), Oral Trail-Making, Verbal Fluency by category and letter, Number Span, and Digit Ordering Test

-- primary outcome: global cognition composite from the above tests with 3 years of CE use vs placebo

-- prespecified secondary endpoint: change in the composite with 3 years of MVM supplementation

-- treatment effects were also examined for executive function and memory composite scores, as well as pre-specified subgroups at higher risk for cognitive decline (see specifics below)

    -- executive function includes attentional control, working memory, inhibition, and problem-solving (many of these skills originate in the prefrontal cortex)

-- tertiary outcomes included the Cognitive Change Index (a self-report measure of cognitive concerns) and the 15-item version of the Geriatric Depression Scale

 

Results:

-- medication adherence (at least 75% of meds taken) decreased from 92% at year 1 to 84% at year 3

-- CE had no effect on global cognition for any of the analyses, and did not add anything to MVM: mean z-score 0.03 (-0.02 to 0.08), p=0.28

    -- z-score is the number of standard deviations from the mean value of the reference population

-- MVM vs placebo: mean z-score 0.07 (0.02-0.12), p=0.007, statistically significant

-- Baseline MVM usage was not associated with MVM response in the study:

    -- no prior use: z-score 0.062 (-0.002 to 0.13)

    -- prior use: 0.77 z-score (0.003 to 0.15)

        -- interaction between the two: p=0.88

These graphs show the change in global cognition relative to baseline over the 3 years, for CE (on left) and MVM (on right)

 

-- cardiovascular disease:

    -- those with a history of CVD: z-score 0.14 (-0.02 to 0.31)

    -- those without a history of CVD: z-score 0.06 (0.01-0.11), with nominal p=0.01 for interaction

 But, those with no CVD history overall started at a higher global cognition composite and improved to a higher level at 3 years than those with CVD history

  

--review of subgroups:

    -- pretty much all of them at least trended to MVM benefit over placebo

    -- though the interaction between the groupings (eg hypertension or not) was not significantly different, there was a significant individual difference by:

        -- age >70

        -- females

        -- upper half of baseline cognitive function

        -- lower half of depressive symptoms

        -- no CVD at baseline (though, as mentioned, the relative increase in those with CVD was higher, but very wide confidence intervals)

        -- hypertension

        -- no diabetes

 

-- other measures (only MVM had significant results):

    -- memory: improved, with z-score 0.06 (0.002-0.13), p=0.04

    -- executive function: improved, with z-score 0.06 (0.01-0.11), p=0.02

 

Commentary:

-- cognitive dysfunction is common (>46 million people worldwide), functionally impairing (and scary…), and without much evidence of benefit from any meds in adults with mild cognitive impairment due to Alzheimer’s

    -- the Alzheimer's foundation notes that Alzheimer's disease affects >6 million cases in the US and is the 6^th leading cause of death in adults

-- cocoa seemed like a reasonable possibility for a nonpharmacologic intervention: has flavonoids, theobromine and caffeine; the flavonoids cross the blood-brain barrier, are detected in the brain, seem to improved cerebral vasodilation, may also increase blood flow/neurogenesis in the hippocampus (important for lerning and memory) and even disrupt amyloid beta aggregation (see https://pubmed.ncbi.nlm.nih.gov/21982844/ ). And several observational assessments have suggested benefit of cocoa (though maybe smarter people eat more chocolate, biasing the results???)

-- many of the previous studies of supplements on cognitive function are either short-term or had few participants. This COSMOS-m study may have had different results in part because they were assessing only long-term cognitive outcomes measured (at least 1 year)

    -- the only long-term other study was the 12-year Physician’s Health Study II of older US male physicians >65yo (quite a specific group, defined by sex, being healthier than the general population and having higher education level, all limiting generalizability); and the Physician's study also had less challenging cognitive testing than COSMOS-m, used a different vitamin/mineral combination, and did not check cognitive testing for 2.5 years after starting the MVM (thereby potentially missing finding earlier changes)

-- there are several micronutrients and minerals that play an important role in normal brain function, and deficiencies of some are associated with increased cognitive decline. Studies of individual micronutrients have had disappointing results, though, I would argue, healthy foods have a complex assortment of micronutrients, there may well be important combinations and interactions between them in the human body, and looking at the single individual one is basically reductionist and does not necessarily mean that that micronutrient is unnecessary…

 

-- this COSMOS-m study found that MVM but not CE led to improved global cognition, memory and executive function

-- those with CVD had more of an improvement, but those without CVD still had improvement, and baseline cognitive function in those without CVD and their improvement on MVM were higher than those with CVD (sort of similar to statins: primary prevention and secondary prevention have about the same 30% decrease in cardiovascular events, but those with secondary prevention, ie having CVD, still have a higher absolute risk of an ensuing event)

    -- there are studies finding that those with CVD have lower micronutrient levels, supporting the use of MVM in those deficient

    -- it is notable that much of the pretty dramatic improvement in those with baseline CVD is related to the significantly decreased cognition in those on placebo.  ??is this related to progression of their atherosclerosis intracerebrally? or to increased incidence of atrial fibrillation (similar risk factors) leading to decreased cognition (see http://gmodestmedblogs.blogspot.com/2022/09/afib-anticoag-decreases-dementia.html)? and would having more people on statins help (after all, essentially everyone in this group had mean age 73 and many comorbidities that would qualify for statins, though only 44% in this cohort were on statins)?? perhaps it would be more effective to use statins than MVMs???

-- as an estimate of the clinical significance of the MVM improvement, MVM (vs placebo) had a treatment effect of 0.083 standard deviations at year 3, which translates to baseline scores of individuals who are 1.8 years apart in age (ie MVM appears to have slowed cognitive aging by about 1.8 years)

-- it should be noted that the overall medical literature does not support multivitamin usage: the USPSTF just released (06/21/2022) their recommendation that "the current evidence is insufficient to assess the balance of benefits and harms of the use of multivitamin supplements for the prevention of cardiovascular disease or cancer" (https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/vitamin-supplementation-to-prevent-cvd-and-cancer-preventive-medication )

    -- but, there was no assessment of cognitive function as part of the review for this recommendation (though, of note, they do cite the COSMOS study as the largest trial finding no effect of these supplements, but that trial was the initial one assessing all-cause mortality or cardiovascular/cancer outcomes, and they also note that the followup period of 3.6 years may have been too short to assess these outcomes. 

    -- and, these USPSTF recommendations precede this current COSMOS-m global cognition analysis

  

Limitations:

-- this study was an "ancillary study" of the initial COSMOS trial, and, as such, does not have the statistical rigor of a full-fledged RCT assessing cognitive outcomes

-- of those recruited into the study, 92% completed at least one cognitive assessment at 1 year (only 77% completed all assessments), and those missing at least one assessment were more likely to be from underrepresented racial/ethnic groups, had less physical activity and chocolate intake, and were more likely to be smoking, have lower education and a lower baseline Telephone Interview of Cognitive Status scores at baseline, skewing the population assessed (already pretty skewed, with 89% white) and potentially limiting generalizability of the results to the general population

-- almost half the people were on baseline multivitamins, and medication adherence decreased to 84% by the end of the study, limiting assessing the potential benefit of MVM itself in preventing cognitive decline (though post hoc analysis suggested no difference in response if on MVM at trial start)

-- all clinical conditions, adherence to meds,  and alcohol/smoking/chocolate were by self-report, raising the possibility of inaccuracies

-- the subgroup findings are of interest but should be treated with a grain of salt: these were not prespecified outcomes and thereby do not have the same statistical rigor. For example, the CVD group had more men who tended to be older, have higher BMI, more hypertension, more statin use, more depression, less physical activity, and lower Telephone Interview of Cognitive Status score: baseline differences that themselves might have altered the results

-- as with all composite outcome assessments, the summation of tests may not reflect the real, clinical importance of the testing: not all tests elicit equal assessment of important cognitive ability, and, for example, equating the value of 10-minute recall with that of digit ordering may distort the results (ie, the former may be way more predictive of future cognitive status than the latter, and just adding them together as numeric equals may lead to erroneous results). This is similar to many of the cardiovascular outcomes in other studies: adding together stroke with getting coronary revascularization in a composite measurement of benefit may undercut the fact that stroke may well be a much more serious/disabling condition than having revascularization (i personally would prefer the latter, given the choice....)

-- we do not have cutpoints for the different cognitive tests to know what difference is actually clinically meaningful. Their modeling did suggest that MVM decreases the aging effect by 1.8 years. But many metrics (eg, for depression, physical functioning) do have what are considered “clinically meaningful differences”, which are not known for many of the tests done in the study

-- we do not have granular data on diets to know if there is some threshold of “eating healthy”, perhaps using the Mediterranean diet as a marker, that obviates the need for MVM to improve or maintain cognitive capacity, or if changing diet itself would have the same (or more) effect than taking a pill. And perhaps a combination of diet and exercise might help.  There are studies suggesting that cognitive health is better with diet and exercise (eg see http://gmodestmedblogs.blogspot.com/2019/07/dementia-genetics-and-lifestyle-both.html)

-- no individual-level data on genetic markers predisposing to Alzheimer’s

 

So, pretty interesting study suggesting that multivitamin supplement seemed to improve global cognition, with a few points:

-- the quality of the cognitive testing done in this COSMOS-m study is really impressive, with more aggressive testing than usually performed, which increases the likelihood of picking up smaller degrees of cognitive decline as well as cognitive decline in those functioning at a higher level

-- the quantity of different tests was also impressive and assessed many domains of cognitive function

-- so this study does open the door to a potential large benefit from a standard, over-the-counter, relatively inexpensive multivitamin preparation without evidence of any significant harms (perhaps some GI adverse effects from the iron) to improve cognition (a really disabling condition)

-- it would have been helpful to know a few more specifics, such as whether vitamins helped in those eating healthy, Mediterranean-type diets... 

    -- as with many intriguing studies, this study does raise more questions than answers

    -- which leads to the typical mantra: "more studies are needed".... (apparently, these authors in COSMOS are planning a larger trial of a more diverse population in the future, assessing blood biomarkers, brain imaging, and microbiome changes, per STAT news (see https://www.statnews.com/2022/09/14/daily-multivitamins-improved-the-brain-function-of-seniors-in-a-trial-but-plenty-of-questions-remain/ )

geoff

 

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@uphams.org

 

to get access to all of the blogs (2 options):

1. go to http://gmodestmedblogs.blogspot.com/ to see them in reverse chronological order

2. click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

3. or you can just click on the magnifying glass on top right, then  type in a name in the search box and get all the blogs with that name in them

 

or: go to https://www.bucommunitymedicine.org/ , a website from the Community Medicine section at Boston Medical Center.  This site does have a very searchable and accessible list of my blogs (though there have been a few that did not upload over the last year or two). but overall it is much easier to view blogs and displays more at a time.

 

 

please feel free to circulate this to others. also, if you send me their emails, i can add them to the list