Meta-Analyses

Positive Studies

Etgen et al evaluated the relationship between Vitamin D and cognitive impairment and dementia through a meta-analysis of case control, cross-sectional, longitudinal, prospective and interventional trials (Etgen et al., 2012). Five cross-sectional studies including 5,686 participants were included. Lower Vitamin D concentrations had significantly increased risk of cognitive impairment, OR 2.37 (1.77-3.17, 95% CI) p= <0.0001. There was significant heterogeneity. Only two studies were included in the longitudinal meta-analysis. Vitamin D deficiency was associated increased risk of cognitive impairment at follow up, OR 2.49 (1.74-3.56, 95% CI) p= <0.001.

A further meta-analysis from 2012 suggests lower concentrations of Vitamin D are associated with poor cognition and dementia (Balion et al., 2012). A total of 37 studies were included in the analysis, from cross-sectional to RCTs, with sample sizes varying from 27 to 17,099. MMSE was the most frequently used cognitive test. In the meta- analysis of Vitamin D and AD versus control groups, six studies with 888 participants were included. The results showed those with AD had lower concentrations of serum 25(OH)D compared with controls, OR -6.2 (-10.6 - -1.8, 95% CI). In further analysis comparing concentrations of 25(OH)D >50nmol/L and <50nmol/L with MMSE scores from eight studies with 2,749 participants, analysis showed higher Vitamin D levels associated with higher MMSE scores, OR 1.2 (0.5 – 1.9, 95% CI).

In 2013, Zhao et al published a meta-analysis looking at the association of Vitamin D deficiency, AD and PD (Zhao et al., 2013). Six AD studies met the eligibility criteria with included studies comparing people with AD or PD with healthy controls. Results showed that in the overall analysis, those with AD were more likely to have lower 25(OH)D levels. However there was significant heterogeneity between all the studies included in the analysis. Five studies on PD met criteria and again analysis showed that those with PD were more likely to have lower Vitamin D levels. Once again there was significant heterogeneity between the studies included.

dementia and Alzheimer’s Disease (Shen and Ji, 2015). Of the five studies, two were prospective cohort studies and three were cross-sectional. In terms of Vitamin D deficiency and AD, three studies were included in the analysis, two prospective and one cross sectional. There was increased risk of AD in those subjects with Vitamin D levels <50nmol/L, OR 1.21 (1.01-1.40, 95% CI). There was also an association between Vitamin D deficiency (<50nmol/L) and development of dementia, OR 1.63 (1.09-2.16, 95% CI). No heterogeneity was noted.

Annweiler et al performed a meta-analysis of studies evaluating the relationship between Vitamin D and cognitive impairment in Asians and included seven studies with 1,179 participants (Annweiler et al., 2016). The mean difference between 25(OH)D between cognitively impaired and cognitively normal individuals was -17nmol/L (-28.4 - -5.7, 95%CI). A subgroup analysis based on AD as the cognitive outcome showed a mean 25(OH)D difference of -37.2nmol/L (-39.3 - -34.9, 95% CI).

A more recent meta-analysis reports an association between Vitamin D deficiency and increased dementia risk in longitudinal, prospective, and RCT studies (Sommer et al., 2017). The authors included five articles in the analysis with a total of 18,933 participants and found that Vitamin D deficiency defined as <25nmol/L was associated with higher risk of dementia, OR 1.54 (1.19-1.99, 95% CI) with low heterogeneity reported.

Goodwill et al report from their meta-analysis that observational data points to an association between Vitamin D and cognition but interventional studies are as yet to show convincing evidence of a benefit from supplementation with Vitamin D (Goodwill and Szoeke, 2017). This meta-analysis was performed using 26 observational studies and three interventional studies. Low Vitamin D levels were associated with poorer cognitive scores, OR 1.24 (1.14-1.35, 95% CI) and cognitive decline, OR 1.26 (1.09-1.22, 95% CI). In intervention studies, Vitamin D supplementation yielded no benefit compared with controls.

A recent meta-analysis form 2018 found an association between Vitamin D concentrations and dementia (Jayedi et al., 2018). This meta-analysis included retrospective and prospective cohort studies and excluded case control studies, studies with only two categories of Vitamin D concentrations, and studies conducted in populations with specific diseases. A total of eight studies were included with 28,354 participants with 1,953 cases of dementia and 1,607 cases of dementia with a median follow-up time ranging from 5 to 21 years. In the seven studies analysing relationship between Vitamin D and dementia, Vitamin D concentrations of 25-50nmol/L did not appear to be associated with risk of dementia, pooled HR 1.09 (0.95-1.24, 95% CI). Vitamin D concentration <25nmol/L was associated with risk of dementia, pooled HR 1.33 (1.08-1.58, 95% CI). There appeared to be a decreased risk of dementia with higher concentrations of 25(OH)D, with a pooled HR 0.83 (0.70-0.96, 95% CI) with 25nmol/L increment in serum 25(OH)D concentrations, but in the context of marked heterogeneity of individual studies. Six studies evaluated the relationship between Vitamin D and risk of AD. In those with Vitamin D concentrations of 25-50nmol/L were not associated with risk of AD, pooled HR 1.19 (0.96-1.4, 95% CI) or with concentrations <25nmol/L, HR 1.31 (0.98-1.675, 95% CI). There appeared to be a dose response relationship with a U shaped curve for Vitamin D and risk of dementia, and a more linear relationship for risk of AD.

7.13 Discussion

There have been many studies investigating the relationship between Vitamin D and cognitive function. A number of studies have shown evidence of a link between Vitamin D deficiency and poor cognition but this link appears to be more consistent in overall test scores and domains related to executive function and attention rather than memory subsets.

Difficulties in comparing these studies include the heterogeneity of the study populations, different definitions of MCI and dementia applied, different assays for Vitamin D and definitions of deficiency and sufficiency used. Many studies also used very crude measures for cognitive assessment.

vascular effects, it is possible its effects are multifactorial.

Ultimately a large placebo controlled RCT is required to determine the effects of Vitamin D on cognition and cognitive decline, using a comprehensive and detailed neuropsychological assessment battery with sufficient follow up period and adequate dosing of Vitamin D to assess the effects of Vitamin D repletion in cognition.

Observational studies still have value in identifying possible confounders and effect modifiers.