ABSTRACT
BACKGROUND/OBJECTIVE: The association between vitamin D levels and cognitive function remains a controversial issue especially in the very old, highly comorbid patients. We address the relative contribution of vitamin D when taking into account potential confounders well known to be involved in cognitive decline. METHODS: We investigated, in a prospective study of 428 very old inpatients from the Geneva geriatric hospital (mean age 85.2 y ± 6.8; 74.1% women), the association between 25-hydroxyvitamin D (25(OH)D) levels with dementia and mild cognitive impairment (MCI), taking into account comorbid conditions, functionality, malnutrition, ApoE genotype, vitamin B12, calcium, and albumin as independent variables. RESULTS: 25(OH)D level was not different between the cognitively normal (n = 200), MCI (n = 46), and demented (n = 182) patients nor between the different etiologies of dementia. In polytomous logistic regression, the 25(OH)D level neither as a continuous variable nor as a categorical variable increased the risk to be MCI or demented. The 25(OH)D level did not predict conversion from normal or MCI to dementia in the 315 subjects who completed the 2 years follow-up. No changes were observed in the full adjusted model after taking into account the independent variables. Similarly, considering only the group of cognitively normal subjects, the 25(OH)D level was not associated with impairment of specific cognitive domain. CONCLUSION: In this cohort of old hospitalized patients with a high burden of comorbidities, the 25(OH)D level alone or adjusted for confounders is not associated with cognitive status and did not predict conversion to dementia.
