|dc.description.abstract||Objective: The association between serum 25-hydroxyvitamin D (s25OHD) concentrations and cognitive function has been investigated in only a few studies of middle aged adults (40-60 years); with inconsistent results found. The main objective of this Canterbury-based study was to examine the association between s25OHD levels and mild cognitive impairment (MCI) in participants in their early fifties, while adjusting for a number of confounders. The secondary objective was to assess if this association would differ when stratified by gender.
Design: Cross-sectional study participants, 366 adults aged 49-53 years living within the Canterbury District Health Board region from the Canterbury Health, Aging and Life Course study, were divided into two cognitive function groups: MCI and normal cognitive function (NCF). A score of <26 on the Montreal Cognitive Assessment test was defined as MCI and NCF was defined as a score ≥26 out of 30. Logistic regression was used to investigate the relationship between the cognitive function groups (outcome variable) and two s25OHD groups: insufficient s25OHD (≤50 nanomoles per litre (nmol/L)) and sufficient s25OHD (>50 nmol/L). Age, gender, education, ethnicity, annual household income, body mass index, depression, season of blood draw, alcohol intake, current smoking status, serum intact parathyroid hormone, serum creatinine, general health, vitamin D food intake, calcium food intake, vitamin D supplement use and chronic diseases were considered as potential confounders.
Results: In the univariate and multivariate logistic regression with cognitive function as the binary outcome variable (MCI (n=89) and NCF (n=277)) and vitamin D as a continuous predictor variable, the results did not reach statistical significance (crude odds ratio (OR) 1 (0.99, 1.01) p=0.306, adjusted OR 0.99 (0.98, 1.00) p=0.084). However, an inverse association between s25OHD levels and cognitive function were observed in the univariate and multivariate logistic regression with vitamin D as a categorical variable. Compared to those with MCI, participants with NCF belonged more often to the insufficient s25OHD group (≤50 nmol/L, n=124) than the sufficient s25OHD group (>50 nmol/L, (n=242)) (crude OR 0.52 (0.30, 0.89) p=0.020, adjusted OR 0.44 (0.24, 0.79) p=0.008). In the analyses stratified by gender, the effect observed was similar, however, the association reached statistical significance only for female participants (adjusted OR 0.43 (0.18, 0.94) p=0.041) and not for male participants (adjusted OR 0.48 (0.18, 1.19) p=0.127). Further adjusted for vitamin D supplement use and season, did not change the effect size, however was just statistically significant, for female participants (adjusted OR 0.43 (0.17, 0.99) p=0.053).
Conclusions: Those with MCI belonged more often to the s25OHD group >50 nmol/L than the s25OHD group ≤50 nmol/L in a middle aged community-dwelling cohort, mostly for women, and appeared to be due to vitamin D supplement use. However, causality cannot be concluded due to the cross-sectional design of this study.||