Study at a glance
What did studies compare about vitamin D, thinking skills and dementia?
This page compares studies linking vitamin D blood levels with later dementia or changes in thinking skills, and trials testing vitamin D3 supplements. A new dementia diagnosis and a change on a memory or planning test are different outcomes. A link with low vitamin D does not by itself show that raising vitamin D will prevent either outcome.
Were low vitamin D levels linked with dementia or declining thinking skills?
One study followed 1,658 older adults initially without dementia and found more later dementia among those with lower vitamin D. Another followed thinking skills in an older group that included people with and without impairment. Lower vitamin D was linked to faster decline in some memory and planning abilities, but not every tested area. Health, behaviour and early disease could influence both vitamin D levels and these outcomes.
Did vitamin D supplements help thinking skills in the trials?
Tests of thinking skills in a trial called VITAL did not show an overall benefit from vitamin D3 over about two to three years. A possible subgroup finding was not a basis for treatment advice based on race. The separate VitaMIND trial assigned 620 adults with age-related decline in thinking skills to vitamin D3 or placebo for two years. It also found no significant benefit on its main outcome or its main additional outcomes. Low vitamin D was predicted by questionnaire, not confirmed by blood tests in every participant.
Do the studies establish a way to prevent dementia?
Both trials had missing later assessments. VITAL included many people with relatively adequate vitamin D levels, so these results do not settle every question about severe deficiency. The studies that observed existing differences cannot separate all other influences or exclude early disease lowering vitamin D. Together, the evidence identifies low vitamin D as a possible sign of risk, but does not establish a supplement strategy, dose or blood target that prevents dementia.
Read the full article below ↓Lower blood vitamin D concentrations have been associated with later dementia and faster decline in some cognitive abilities. That observation matters, but it does not establish that raising the concentration prevents either outcome. This appraisal compares two prospective cohorts and uses randomised trials to test the treatment interpretation. The distinction between a risk marker and a modifiable cause remains central throughout. For the shorter account, read Vitamin D Status, Cognitive Decline and Dementia Associations: Study Summary.
The question changes when the outcome changes
The brain overview provides physiological context. Here the evidence concerns people and measured cognitive outcomes. Incident dementia means a new diagnosis during follow-up. A change in memory or executive-function test scores means a change on particular tasks. Neither is identical to a laboratory marker of neuronal injury, and a statistical association with one outcome should not automatically be described as an association with all neurological disease.
Executive functions help people organise actions, switch between tasks and hold information in mind. Episodic memory concerns the recall of experiences or learned material. Tests sample aspects of these abilities under specified conditions. Their results can also be influenced by education, language, sensory difficulties, mood and familiarity with testing. A useful study identifies the domain, instrument and timing rather than reporting an undifferentiated improvement or deterioration in brain health.
A prospective dementia cohort
Littlejohns and colleagues analysed 1,658 older adults from the Cardiovascular Health Study, initially free of dementia, stroke and cardiovascular disease. Baseline serum 25-hydroxyvitamin D, the usual status marker, was measured using liquid chromatography: mass spectrometry. Dementia was adjudicated using clinical and other information. During an average 5.6 years, 171 participants developed dementia, including 102 classified as Alzheimer disease. These are overlapping outcomes, not two independent sets of events.
Compared with concentrations of at least 50 nanomoles per litre, the fully adjusted dementia hazard ratio was 1.53 for 25 to below 50 and 2.25 for below 25. The corresponding 95% confidence intervals were 1.06 to 2.21 and 1.23 to 4.13. A hazard ratio describes relative event rates over follow-up; it is not the percentage of people who will develop dementia, nor the reduction expected from a supplement.
The fully adjusted model included age, season of blood collection, education, sex, body mass index, smoking, alcohol consumption and depressive symptoms. Excluding diagnoses during the first year left the association broadly similar. That helps address one form of reverse causation, but early disease can influence activity, diet and health for longer than one year. A prospective design establishes that the blood sample preceded the recorded diagnosis; it does not guarantee that it preceded the underlying disease process.
Selection also matters. People with unavailable serum or missing follow-up were excluded, and the analysed group omitted those with baseline vascular disease. The lowest vitamin D category was relatively small, contributing to wide confidence intervals. Adjustment for measured variables cannot repair every difference between included and excluded participants. The results should therefore be interpreted within this selected older population rather than treated as a universal prediction rule.
A cohort with repeated cognitive testing
Miller and colleagues studied a multiethnic older cohort recruited through community and memory-clinic routes. There were 382 participants at baseline; 318 contributed longitudinal assessments over an average 4.8 years, with smaller complete-data samples for individual models. Unlike the dementia-free cohort, this population included people with normal cognition, mild impairment and dementia at entry. Its principal longitudinal question concerned trajectories measured with the Spanish and English Neuropsychological Assessment Scales. Adjusted models included age, sex, race or ethnicity, education, body mass index, season, vascular risk and APOE4 genotype, a genetic risk factor.
Lower vitamin D categories were associated with faster adjusted decline in episodic memory and executive function. For the deficient category, the additional annual decline relative to the adequate category was approximately 0.04 standardised units for episodic memory and 0.05 for executive function. Both estimates had a standard error of 0.02; the corresponding P values were 0.049 and 0.01. The memory comparison was therefore close to the conventional significance boundary. The categorical comparisons did not show significant differences for every domain. Continuous and categorical analyses also gave somewhat different patterns, illustrating how exposure modelling affects interpretation.
A standardised unit expresses a difference relative to a score distribution. It is not a year of ageing, a diagnosis or a direct measure of everyday independence. Translating a slope into a clinical benefit requires evidence that changing the exposure changes the trajectory and that the difference is meaningful to patients. Multiplying a statistical coefficient across many years without checking the model assumptions can create an unjustified impression of certainty.
The cohort lacked direct measures of some relevant behaviours, including diet, sun exposure and exercise. Participants lost from longitudinal assessment were not necessarily comparable to those retained. Removing people who already had dementia did not eliminate the reported association, but residual confounding and selection remained possible. The study supports further investigation of the marker; it cannot by itself determine whether correcting it alters cognition.
The two cohorts do not define one treatment threshold
Their vitamin D categories were not identical. Littlejohns used nanomoles per litre, with a lowest category below 25; Miller used nanograms per millilitre, with a deficient category below 12 and an adequate category from 20 to below 50. Unit conversion is necessary before comparing boundaries. Even after conversion, differences in assays, recruitment, baseline health and outcome definitions remain. Matching category labels is not enough to establish that two studies tested the same exposure contrast.
This is one reason to distinguish vitamin D status from vitamin D effect. A circulating concentration can track several influences at once. It may be a contributor to an outcome, a consequence of other health changes, a marker of behaviour, or some combination. Observational associations do not identify the benefit of a particular dose, the time at which treatment would need to begin or a blood target for preventing dementia.
Randomisation tests a different proposition
The VITAL cognitive report by Kang and colleagues examined two separate participant subsets within the same parent trial. It included 3,424 people assessed by telephone and 794 assessed in person. They should not be counted as two independent trials of the underlying intervention. Participants were assigned vitamin D3 at 2,000 international units daily or placebo within a factorial trial that also assigned omega-3 treatment.
Over approximately two to three years, the pooled difference in annual change in the primary global cognitive score was 0.01 standardised units, with a 95% confidence interval from minus 0.01 to 0.02. This did not demonstrate a benefit. The trial also found no overall benefit on its key secondary cognitive outcomes. Many participants had relatively adequate baseline status, which limits how directly the findings answer questions about people with severe deficiency.
A prespecified analysis suggested a more favourable global-score result among Black participants, while baseline vitamin D concentration did not significantly modify the effect. The subgroup comparisons were not all adjusted for multiplicity. They require replication and should not become a race-based treatment recommendation. Randomisation strengthens the overall treatment comparison, but a favourable subgroup result can still arise by chance, particularly when several interactions are examined.
Follow-up structure deserves attention as well. Not every participant completed both cognitive assessments, and additional Black participants entered the telephone assessment programme at its later wave. The investigators used longitudinal models and sensitivity analyses to examine these issues. Model-based use of incomplete data is reasonable under stated assumptions; it is not the same as observing every person at every scheduled time point.
A more recent trial with cognitive impairment at entry
The 2025 VitaMIND report provides an additional boundary on interpretation. It randomised 620 adults aged at least fifty with age-associated cognitive decline to 4,000 international units of vitamin D3 daily or placebo for twenty-four months. It reported no significant benefit on the primary executive-function outcome or the principal secondary outcomes. These were study exposures, not recommended doses for readers.
Eligibility for low vitamin D status was based on a prediction questionnaire, rather than a blood result in every participant. Only a small subgroup supplied baseline blood samples. Consequently, the trial should not be described as a study in which all 620 participants had laboratory-confirmed deficiency. Completion of the primary test at twenty-four months was also substantially below the number randomised; reported imputation analyses do not make the missing observations disappear.
VitaMIND is useful recent supporting evidence, but this appraisal does not independently reconstruct its inaccessible supplementary prediction material. Its complete principal report was read, including the published methods and tabulated outcomes. The interpretation is confined to the reported overall findings and visible design limitations. It neither proves that every deficient population has been adequately studied nor supports a claim that cognitive benefit from treating severe deficiency has already been established.
Mechanisms remain hypotheses within these human studies
The physiological discussions of neuroendocrine integration, inflammatory signalling and oxidative stress offer possible explanations for an association. These cohorts did not directly demonstrate that any of those pathways mediated their cognitive findings. A plausible mechanism makes a research question coherent, but does not resolve confounding or turn a null supplementation comparison into a positive clinical result.
Similarly, receptor-dependent gene expression in experimental systems should not be used to predict the size of a human treatment effect. Human exposure, tissue response, disease stage and concurrent illness differ from controlled cell conditions. Establishing mediation would require appropriately timed measurements and analyses, ideally supported by interventions that change the proposed pathway. Without that evidence, a mechanistic account should remain clearly labelled as a hypothesis.
What can reasonably be concluded
The observational signal is sufficiently consistent in these cohorts to merit attention, while the treatment proposition remains more restricted. The selected randomised comparisons did not establish an overall cognitive benefit in the populations and time periods tested. That combination is compatible with several explanations: confounding, reverse causation, a critical exposure window, differences in baseline status, or effects too small or delayed for a given study to detect. These possibilities are not equally proven simply because they can be proposed.
The evidence hierarchy is useful when it is applied to a specific question. Cohorts can describe long-term associations and identify populations for further study. Trials can test an assigned intervention, but only within their eligibility, dose, follow-up and outcome boundaries. Neither design should be asked to establish a conclusion it did not measure. The current appraisal supports cautious interpretation of low status as a risk marker, without claiming a proven strategy for preventing dementia.
Future studies would be more informative if they documented baseline status consistently, retained participants across repeated assessments, specified cognitive domains in advance and measured clinically meaningful outcomes. Long follow-up and transparent handling of missing data are especially important when disease develops slowly. Any subgroup claim needs an interaction analysis and independent confirmation. These requirements keep the next research question concrete: which intervention, in which population, changes which outcome, over what period?
Scientific sources
Littlejohns TJ, Henley WE, Lang IA, et al. Vitamin D and the risk of dementia and Alzheimer disease. Neurology. 2014;83:920 to 928. DOI:10.1212/WNL.0000000000000755. PMID:25098535. Complete principal paper, two figures and three tables examined; supplementary tables were inaccessible.
Miller JW, Harvey DJ, Beckett LA, et al. Vitamin D Status and Rates of Cognitive Decline in a Multiethnic Cohort of Older Adults. JAMA Neurology. 2015;72:1295 to 1303. DOI:10.1001/jamaneurol.2015.2115. PMID:26366714. Complete principal paper, trajectory figure and four tables examined; supplemental table was inaccessible.
Kang JH, Vyas CM, Okereke OI, et al. Effect of vitamin D on cognitive decline: results from two ancillary studies of the VITAL randomized trial. Scientific Reports. 2021;11:23253. DOI:10.1038/s41598-021-02485-8. PMID:34853363. Complete principal paper, main figure, six tables and eight-page supplementary appendix examined. A validation plot cited in the appendix was not present in the retrieved file.
Corbett A, Taylor R, Llewellyn D, et al. Impact of Vitamin D Supplementation on Cognition in Adults With Mild to Moderate Vitamin D Deficiency: Outcomes From the VitaMIND Randomized Controlled Trial. Journal of the American Medical Directors Association. 2025;26:105711. DOI:10.1016/j.jamda.2025.105711. PMID:40480279. Complete principal-report text and tabulated outcomes read; supplementary prediction material was inaccessible. Used as bounded supporting evidence.
Related reading
Current Evidence on the Association of Micronutrient Malnutrition with Mild Cognitive Impairment, Frailty, and Cognitive Frailty among Older Adults: A Scoping Review
The purpose of this scoping review was to map current evidence on the association between micronutrient deficiency in blood and mild cognitive impairment, frailty, and cognitive frailty among older adults.
Original paper ↗Read more →Dietary vitamin D intake and 2-year changes in cognitive function in older adults with overweight or obesity and metabolic syndrome
We aimed to evaluate the association between dietary vitamin D intake and 2-year cognitive changes in older adults at risk of cognitive decline.
Original paper ↗Read more →Diversity and inclusivity in Australian dementia prevention research: A mixed methods review
Using co‐creation methodologies, the workshop aimed to generate consensus‐based recommendations for governments, institutions, and individual researchers to improve inclusivity and representativeness in future dementia prevention research.
Original paper ↗Read more →Factors associated with cognitive function outcomes among older adults in Kuwait: A cross-sectional study
A normal MMSE score was reported for 51.0% (n = 129) of the sample, with 34.7% and 14.2% of participants having mild and moderate/severe cognitive impairment, respectively.
Original paper ↗Read more →Impact of Vitamin D Status and Supplementation on Brain-Derived Neurotrophic Factor and Mood-Cognitive Outcomes: A Structured Narrative Review
Background/Objectives: Vitamin D deficiency is prevalent in higher-latitude regions and among older adults, and has been linked to depressive symptoms and cognitive decline, although the neurobiological link remains unclear.
Original paper ↗Read more →Incidence and Predictors of Cognitive Frailty Among Older Adults: A Community-based Longitudinal Study
(1) Background: Cognitive frailty (CF) is the simultaneous presence of physical frailty and cognitive impairment with an increased risk of dementia.
Original paper ↗Read more →Increased risk of incident dementia associated with vitamin D deficiency in glaucoma patients: a TriNetX cohort study
This study investigated the longitudinal association between VDD and the 5-year incidence of neurodegenerative outcomes in this glaucoma population.
Original paper ↗Read more →Interaction between ω-6 fatty acids intake and blood cadmium on the risk of low cognitive performance in older adults from National Health and Nutrition Examination Survey (NHANES) 2011 to 2014
This study evaluated the interaction effect between blood cadmium and ω-6 fatty acids intake on low cognitive performance in Americans.
Original paper ↗Read more →Interpretable machine learning models for predicting cognitive impairment using NHANES neuropsychological tests: nutritional and sociodemographic associations
Early identification of individuals at risk for cognitive impairment is essential for timely intervention and public health planning.
Original paper ↗Read more →Investigating Modifiable Factors Associated with Cognitive Decline: Insights from the UK Biobank
After excluding participants with dementia and substantial missing data, 453,950 individuals from UK Biobank (UKB) were included.
Original paper ↗Read more →Low Vitamin D and Its Association with Cognitive Impairment and Dementia
Cross-sectional studies have consistently found that vitamin D levels are significantly low in individuals with Alzheimer’s disease and cognitive impairment compared to healthy adults.
Original paper ↗Read more →Low vitamin D status and 10-year dementia risk in sensory-impaired adults: a propensity score-matched cohort study
Second, we evaluated healthcare visits as a proxy for detection bias.
Original paper ↗Read more →Prospective longitudinal study of frailty transitions in a community-dwelling cohort of older adults with cognitive impairment
We thus aim to study frailty transitions and change in cognitive status over 1-year follow-up among subjects with cognitive impairment attending a tertiary Memory Clinic.
Original paper ↗Read more →Role of Vitamin D in Cognitive Dysfunction: New Molecular Concepts and Discrepancies between Animal and Human Findings
The developmental disabilities induced by vitamin D deficiency (VDD) include neurological disorders (e.g., attention deficit hyperactivity disorder, autism spectrum disorder, schizophrenia) characterized by cognitive dysfunction.
Original paper ↗Read more →Serum 25-hydroxyvitamin D and cognitive decline in the very old: the Newcastle 85+ Study
Studies investigating the association between 25-hydroxyvitamin D [25(OH)D] and cognition in the very old (85+) are lacking.
Original paper ↗Read more →Serum 25-hydroxyvitamin D and cognitive performance in hospitalized older adults: a cross-sectional study from Central Poland
This study aimed to investigate the association between serum 25(OH)D concentration and cognitive performance, as well as the odds of cognitive impairment and documented dementia in a hospitalized geriatric population.
Original paper ↗Read more →Study of Diet Habits and Cognitive Function in the Chinese Middle-Aged and Elderly Population: The Association between Folic Acid, B Vitamins, Vitamin D, Coenzyme Q10 Supplementation and Cognitive Ability
The objective of the present cross-sectional study was to evaluate the relationship between cognitive ability and folic acid, B vitamins, vitamin D (VD) and Coenzyme Q10 (CoQ10) supplementation.
Original paper ↗Read more →Summarizing attributable factors and evaluating risk of bias of Mendelian randomization studies for Alzheimer’s dementia and cognitive status: a systematic review and meta-analysis
This systematic review aimed to summarize evidence from Mendelian randomization (MR) studies exploring factors attributable to AD and related cognitive status and to assess its credibility.
Original paper ↗Read more →The Association of Thyroid Disease with Risk of Dementia and Cognitive Impairment: A Systematic Review
Background and Objectives: Cognitive impairment is defined as a reduced ability to perform one or more cognitive functions, potentially leading to dementia if the condition worsens.
Original paper ↗Read more →The interactions between vitamin D and neurofilament light chain levels on cognitive domains in bipolar disorder
Longitudinal or randomised controlled studies enrolling patients with various illness durations and mood statuses are required to validate our findings.
Original paper ↗Read more →Vitamin D and Cognitive Impairment
Objective: Therefore, this study aimed to examine the prevalence of vitamin D deficiency in the Thai elderly population and an association between vitamin D status and cognitive function, adiposity, and insulin sensitivity.
Original paper ↗Read more →Original sources
Vitamin D and the risk of dementia and Alzheimer disease
Published in Neurology (2014).
Publication details
Littlejohns TJ, Henley WE, Lang IA, et al. Vitamin D and the risk of dementia and Alzheimer disease. Neurology. 2014;83:920 to 928. DOI:10.1212/WNL.0000000000000755. PMID:25098535. Complete principal paper, two figures and three tables examined; supplementary tables were inaccessible.
Vitamin D Status and Rates of Cognitive Decline in a Multiethnic Cohort of Older Adults
Published in JAMA Neurology (2015).
Publication details
Miller JW, Harvey DJ, Beckett LA, et al. Vitamin D Status and Rates of Cognitive Decline in a Multiethnic Cohort of Older Adults. JAMA Neurology. 2015;72:1295 to 1303. DOI:10.1001/jamaneurol.2015.2115. PMID:26366714. Complete principal paper, trajectory figure and four tables examined; supplemental table was inaccessible.
Effect of vitamin D on cognitive decline: results from two ancillary studies of the VITAL randomized trial
Published in Scientific Reports (2021).
Publication details
Kang JH, Vyas CM, Okereke OI, et al. Effect of vitamin D on cognitive decline: results from two ancillary studies of the VITAL randomized trial. Scientific Reports. 2021;11:23253. DOI:10.1038/s41598-021-02485-8. PMID:34853363. Complete principal paper, main figure, six tables and eight-page supplementary appendix examined. A validation plot cited in the appendix was not present in the retrieved file.
Impact of Vitamin D Supplementation on Cognition in Adults With Mild to Moderate Vitamin D Deficiency: Outcomes From the VitaMIND Randomized Controlled Trial
Published in Journal of the American Medical Directors Association (2025).
Publication details
Corbett A, Taylor R, Llewellyn D, et al. Impact of Vitamin D Supplementation on Cognition in Adults With Mild to Moderate Vitamin D Deficiency: Outcomes From the VitaMIND Randomized Controlled Trial. Journal of the American Medical Directors Association. 2025;26:105711. DOI:10.1016/j.jamda.2025.105711. PMID:40480279. Complete principal-report text and tabulated outcomes read; supplementary prediction material was inaccessible. Used as bounded supporting evidence.