|Home | About | Journals | Submit | Contact Us | Français|
The objective of this study was to examine the association between 25-hydroxyvitamin D, 25(OH)D, and cognitive function.
A cross-sectional investigation of 25(OH)D and cognition was completed in 377 black and 703 non-black (mainly Caucasian) elders (65–99 years) participating in the nutrition and memory in elders study. Participants underwent a comprehensive neuropsychological battery, and 25(OH)D concentrations were obtained.
More than 65% of elders had suboptimal 25(OH)D concentrations (≤20 ng/mL or ≤50 nmol/L). Approximately 18% were deficient in 25(OH)D (<10 ng/mL or <25 nmol/L). After adjusting for age, sex, race, body mass index, education, center, kidney function, seasonality, physical activity, and alcohol use, 25(OH)D was associated with better performance on trails A (β = –0.49, p < .03), trails B (β = –0.73, p < .02), digit symbol (β = 0.19, p < .001), matrix reasoning (β = 0.04, p < .02), and block design (β = 0.07, p < .04) tests. Associations remained after adjustment for homocysteine, apoE4 allele, plasma B vitamins, and multivitamin use (y/n). 25(OH)D concentrations >20 ng/mL were associated with better performance on tests of executive function, including trails A (80.5 vs 95, p < .05), trails B (205s vs 226s, p < .05), matrix reasoning (7.8 vs 7.0, p = .03), and digit symbol (31.5 vs 37, p < .01). There were no associations between 25(OH)D and memory tests. Factor analysis yielded factors for memory, executive function, and attention/processing speed. After adjustment, 25(OH)D was associated with the executive function (β=0.01, p < 0.01) and attention/processing speed factors (β = 0.01, p = .03), but not the memory factor (β = –0.001, p = 0.65).
25(OH)D was positively associated with cognitive performance, particularly with measures of executive function in this elderly population.
VITAMIN D status has implications that extend beyond the well-known effects on bone mineralization. New research suggests that vitamin D has beneficial effects on muscle function (1), cardiovascular health (2), diabetes (3), and cancer prevention (4). More recently, clinical studies have suggested a potential beneficial role of vitamin D for cognitive function (5,6). A growing body of evidence suggests that the active form of vitamin D is involved in development and adult brain function (7,8). Metabolic pathways for vitamin D have been found in the hippocampus and cerebellum (7), suggesting that vitamin D may be active in areas of the brain involved in planning, processing, and formation of new memories.
Dementia is cognitive decline resulting in memory loss and impairment of normal daily function. The pathology of dementia in the elderly is complex and multifactorial. Alzheimer's type dementia, the most common form of age-associated dementia, affects approximately 10% of adults more than the age of 65 years and 47% of adults more than 85 years in the United States (9). Vascular dementia, the second most common form of age-associated dementia, is estimated to account for more than 17% of dementia in the United States (9). Vitamin D may help in the prevention of these neurodegenerative diseases of aging through protection against comorbidities such as cardiovascular and cerebrovascular disease (35), by modulation of oxidation and inflammation (8,10) and by protection of nerve conduction (8,11).
The global population is rapidly aging, and those aged ≥85 years represent the fastest growing segment (12,13). This demographic shift marks an increase in the number of elders requiring homecare services. These elders bear a disproportionate burden of disease and are more likely to have cognitive impairment. Estimates reveal that more than 40% of US and European elderly adults are deficient in vitamin D. Elders requiring homecare are particularly at risk for vitamin D insufficiency because of sunlight deprivation, inadequate nutriture, age-related dermatological changes, and impairments in renal function. These risks may be further exacerbated in certain racial or gender groups due to excess comorbidities, differences in nutritional requirements, and variation in dietary intakes.
Identifying nutritional factors that may mitigate cognitive dysfunction and help preserve independent living has a significant economic and public health benefit. This study was designed to evaluate associations between 25-hydroxyvitamin D, 25(OH)D, concentration and cognitive function, specifically aspects of executive function, memory, and attention/processing speed, in an urban dwelling, racially diverse elderly population in need of homecare services in Boston, MA.
The detailed methodology for the nutrition and memory in elders (NAME) study has been described elsewhere (14). Participants were recruited through four Boston homecare agencies and were eligible if they were at least 60 years of age, English speaking, and free of severe auditory and visual impairment, and specific conditions (epilepsy, human immunodeficiency virus, schizophrenia, bipolar disorder, metal retardation, or brain tumor). All eligible participants provided informed consent, and the protocol and consent forms were approved by the Institutional Review Board at Tufts Medical Center.
A fasting blood sample was obtained, immediately aliquoted, centrifuged, and transported to the laboratory on ice. Circulating 25(OH)D concentrations are considered the best indicator of vitamin D deficiency, insufficiency, hypovitaminosis, sufficiency, and toxicity (15) and were, therefore, utilized in this study. Samples for analysis were stored at –70° C and were analyzed uniformly in batch. Prior to analysis, 50 μL of plasma was extracted from the stored samples, and procedures for evaluation of 25(OH)D via the Diasorin 125I RIA assay were followed (16). Laboratory reference ranges were 8–38 ng/mL, and values outside of this range were rerun for confirmatory analysis. Categories of vitamin D status were defined as deficient (<10 ng/mL), insufficient (10–20 ng/mL), and sufficient (>20 ng/mL; 5,17,18). General information related to demographics and household composition was obtained through an interviewer-administered questionnaire.
Standardized instruments and protocols were used to measure weight and height. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. In cases where standing height was unreliable or unavailable, knee height measures were obtained. In such cases, height was estimated using the prediction equations developed from the National Health and Nutrition Examination Survey (NHANES) III data before BMI was calculated (19). Habitual physical activity was assessed using the Paffenbarger questionnaire (20) and number of miles walked was obtained from the questionnaire and used as a measure of physical activity. This measure was used based on previous reports of a cognitive protective benefit of walking (21). Health history was obtained by self-report and food and nutrient intake, such as information on vitamin E, carotenoids, and omega-3 fatty acids, was captured with a validated semi-quantitative food frequency questionnaire (FFQ; 22). Use of multivitamins or single-nutrient supplements and their frequency were also captured and interviewer verified.
Participants were classified as having diabetes according to the guidelines from the American Diabetes Association (fasting plasma glucose of >126 mg/dl or are using prescribed oral medication for diabetes or insulin). Hypertension (based on the average of the two blood pressure measurements taken during the home visit with an Omron manual inflation blood pressure monitor) was defined as higher ≥140 or 90 mmHg for systolic or diastolic blood pressure, respectively. Kidney function was assessed using the modification of diet in renal disease study group (isotope dilution mass spectroscopy traceable calibration, MDRDidms) equation (23). After indirect calibration of serum creatinine values, glomerular filtration rate (GFR) was estimated using the four-variable MDRD study equation, reexpressed for an isotope dilution mass spectroscopy standard (24). Participants with estimated GFR <60 mL/minute per 1.73m2 were classified with reduced kidney function (25). Vitamin D was categorized as deficient (<10 ng/mL), insufficient (10–20 ng/mL), and sufficient (>20 ng/mL; 17,18).
The mini mental state examination (MMSE; 26) and the North American Adult Reading Test (NAART; 27) were administered to determine eligibility. The MMSE is a 30-point test designed to assess global cognition and screen for dementia. Participants with an MMSE ≤10 or verbal IQ (estimated with the NAART) <75 were excluded from participation in order to prevent floor effects. The neuropsychological test battery, which is described in detail elsewhere (14), was administered at home by trained research assistants. It was designed to assess a broad range of neurological functioning including verbal intelligence (NAART), supraspan learning (WMS-III word list learning; 28), auditory and visual retention (WMS-III logical memory; 28), executive function/mental processing speed (digit symbol coding, trail-making test, mental alternations, WAIS-III block design), visual construction/fluid reasoning (matrix reasoning; 29), verbal fluency (controlled oral word association; 30), and anxiety (self-rated anxiety scale; 31).
The total sample included 1253 participants (771 Caucasian, 438 black, 11 Hispanic, 11 Asian/Pacific islander, and 22 other). A total of 172 participants without 25(OH)D concentrations and one participant with 25(OH)D >80 ng/mL (>200 nmol/L) were excluded from the analysis. Remaining participants were categorized as black (n=377) or non-black (703), after statistical confirmation that there were no differences within the races categorized in the non-black group (data not shown). The non-black group included 668 (95%) Caucasians, 10 (1.4%) Hispanics, 10 (1.4%) Asians/Pacific islanders, and 16 (2.3%) “other race/ethnicity.”
All analyses were completed using SAS (version 9.1; ©2007, SAS Institute, Cary, NC). Variable distributions were evaluated, and those not normally distributed were log transformed to meet the assumptions of the statistical tests, when appropriate. Initial comparisons on demographic and descriptor variables were conducted across race and sex, due to the known differences in vitamin D status by race and differences in cognitive variables by sex. Models were adjusted for the homecare agency from which the participant was selected, to adjust for possible unknown confounders associated with the geographic location of the agency.
Clinical and demographic characteristics were compared by race with two-tailed t tests for two independent samples and with the χ2 test for differences in proportion. Group means were adjusted and compared with analysis of covariance (ANCOVA), and possible interactions were investigated by including interaction terms in the ANCOVA models. Univariate and multiple regression models were constructed to examine associations of plasma 25(OH)D concentration and cognitive function.
Principal components analysis was used as a data reduction technique for evaluation of neuropsychological tests, to address issues associated with multiple testing. The analysis of cognitive factors was completed with an orthogonal rotational procedure (varimax). The adequacy of the correlation was assessed with Bartlett's test and Kaiser–Meyer–Olkin measure. Factors with eigenvalues >1.0 were selected, and factor loadings were defined by items with factor loading scores >|0.4|. The factor scores were incorporated into regression models to explore the association between physiologic vitamin D concentrations and these dimensions of cognitive functioning with and without control for confounders. In confirmatory analyses, representative tests that were specified from each factor were extracted and further explored as the dependent variable to evaluate the relationship of vitamin D concentrations on specific tests of cognition.
The mean age of the sample was 75.0 ± 8.5 years. Seventy five percent of the participants were women (Table 1). Mean BMI exceeded 30, and more than 25% of the participants had not finished high school. Non-black elders were, on average, significantly older than black elders, had lower BMI, and were more likely to have at least a high school education (each p < .001). Hypertension, diabetes, and kidney disease were prevalent in this sample, at 85%, 36%, and 34%, respectively. Blacks were significantly more likely to have hypertension and diabetes than were non-blacks (each p < .001). Non-black elders were significantly more likely than blacks to use multivitamin supplements (56% vs 48%, p=.02), but use did not differ between men and women. Fewer than 20% of the sample used vitamin D supplements, with no significant difference by race. However, women were more than twice as likely as men to use vitamin D supplements.
Non-black elders had significantly higher 25(OH)D concentrations than did blacks (19 vs 16 ng/mL, p < .001, Table 2). Approximately 50% of the sample had total vitamin D intakes <400 IU/day, and only about 30% of the non-blacks and 22% of the blacks had intakes >600 IU/day. Among non-supplement users, only 10% had vitamin D intakes >400 IU /day and fewer than 5% had intakes >600 IU/day.
More than 65% of the population had physiologic 25(OH)D concentrations considered suboptimal (<20 ng/mL). Approximately 18% were considered deficient (<10 ng/mL), 47% insufficient (10–20 ng/mL), and 35% sufficient (>20 ng/mL). Less than 8% of the sample had 25(OH)D concentrations >30 ng/mL. Blacks were significantly more likely to have 25(OH)D concentrations <10 ng/mL (p < .01) and less likely to have concentrations >20 ng/mL (p < .01; Table 2).
There were no interactions by race in our analysis and therefore the groups were combined and race was adjusted for as a confounder in the associations of 25(OH)D and cognition. The Pearson correlation between 25(OH)D concentrations and the MMSE score approached significance (r=.06, p=.05) whereas the correlations between 25(OH)D concentrations and cognitive tests assessing executive functioning were significant: block design (r=.07, p=.04), digit symbol (r=.12, p < .001), digit span (r = 0.08, p < .01), matrix reasoning (r = 0.08, p <.02), trails A (r = –0.08, p < .03), and trails B (r = –.10, p < .02), after adjusting for age, sex, race, BMI, kidney function, education, seasonality, center, physical activity, and alcohol use (Table 3).
In the regression analysis, 25(OH)D concentrations were associated with significantly better performance on trails A, trails B, digit symbol, matrix reasoning, block design, and digit span tests (each p < .05, Table 3). These results remained after additional adjustment for covariates known to be possibly related to cognitive performance, including homocysteine, apoE4 allele status, plasma B vitamins, and multivitamin use (y/n). Additionally, 25(OH)D concentration >20 ng/mL was associated with better mean scores on tests of executive function, including trails A, trails B, matrix reasoning, and digit span (each p < .05). Concentrations between 10 and 20 ng/mL were associated with better mean scores on tests of executive function, compared with the <10 ng/mL group (trails B, block design, and digit span, Table 4). There were no observed associations between 25(OH)D and tests specifically for memory (β=–0.004, p < .54) and no observed difference in mean scores across 25(OH)D categories for tests of memory (Table 4).
Principal components analysis revealed three distinct factors related to memory, executive function, and attention/processing speed. There were significant associations between 25(OH)D and the executive function and attention/processing speed factors, after covariate adjustment (p < .01). The association with executive function remained significant after further adjustment for plasma B vitamins (vitamins B6, B12, and folate) and homocysteine (Table 3). The association with attention/processing speed remained significant after adjustment for homocysteine and for vitamins B12 and folate. However, adjustment for vitamin B6 attenuated the association. The addition of dietary covariates, as obtained by the FFQ, including vitamin E, carotenoids, and omega-3 fatty acids, as well as chronic conditions such as hypertension and diabetes, did not influence the observed associations between 25(OH)D and executive function or attention processing speed. These covariates were not included in the final models.
In this cross-sectional study of urban-dwelling elders receiving home health services, low 25(OH)D was associated with impairments in cognitive function; further associations between 25(OH)D and both global and specific areas of cognitive function were observed. We found that circulating 25(OH)D concentration was most strongly associated with measures of cognition in the areas of executive functioning (complex cognitive tasks, such as planning, problem solving, or sequencing) and attention processing speed. Similar associations between vitamin D and memory were not seen.
These findings are in agreement with recent studies (5,6). The first, a cross-sectional study of 80 ambulatory elders (40 with mild dementia and 40 nondemented), showed that vitamin D deficiency was associated with poorer performance on the short blessed test (a tests of central processing speed) and the clinical dementia rating sum of box scores (a test of global cognition; 5). Similarly, Pryzbesky and colleagues reported a significant correlation between 25(OH)D concentrations and MMSE score (r=.23, n=32; 6). Consistent with the association between vitamin D insufficiency and cognition, participants with secondary hyperparathyroidism (SHPT, n=21) in the Tromoso study performed worse on tests of working memory, processing speed, and language, than participants without SHPT (n=63), yet no specific associations between 25OHD and cognition were observed (32).
In contrast to our findings, McGrath and colleagues reported no association between 25(OH)D and cognitive function in adolescents (16–19 years) or adults (20–59 years), in the NHANES data, yet the investigators found a very small, but significant inverse association between 25(OH)D and one test of learning and memory in older adults (60–90 years; 33).
Although the earlier studies were limited in sample size or cognitive assessment tools, observations from the NAME study showed positive associations between vitamin D and cognition in a large cohort with a comprehensive neuropsychological battery. Furthermore, the associations observed between vitamin D and cognitive domains associated with subcortical function yield a potential mechanistic role for vitamin D in cognitive function consistent with a potential vasculoprotective role of the hormone. In addition to associations with central nervous system function, vitamin D insufficiency has been associated with increased risk for hypertension (2,34,35), diabetes (3), and impairments in physical function (1,36). Consistently, this population of homebound elders exhibited high prevalence of hypertension (85%) and diabetes (36%). Furthermore, it is known that the above conditions have been associated with an increased risk for cognitive impairment and dementia. However, this cross-sectional study was not designed to further explore the causal determinants of impairment beyond the observed associations.
The pathology of dementia is complex and multifactorial. Evidence supports a role for vitamin D in the protection against neurodegenerative diseases of aging through the inhibition of inducible nitric oxide synthase, an enzyme that is upregulated during ischemic events, Alzheimer's disease, and Parkinson's disease. Vitamin D also upregulates gamma glutamyl transpeptidase activity and subsequently increases glutathione concentrations in glial cells (8), which protect oligodendrocytes and the integrity of the nerve conduction pathway critical to mental processing. Although vitamin D has been shown to protect brain structures through modulation of neuronal calcium homeostasis (11), it may also protect against cognitive impairment, through protection against cardiovascular and cerebrovascular disease secondary to benefits that extend beyond calcium and phosphate homeostasis.
Because elders receiving home care services exhibit inherent limitations in physical function and impairments in daily activities, their accessibility to sunlight sources of vitamin D is generally low. Although the cross-sectional nature of this study limits our ability to show a temporal or causal association between 25(OH)D and cognitive function, a substantial body of evidence suggests that homebound elders may benefit from higher intakes of vitamin D. More than 65% of the population under study had 25(OH)D concentrations less than 20 ng/mL. Although similar results were reported in the wintertime in elders in Boston, MA (37), data for this study were collected throughout all four seasons. Notably, there was no difference in 25(OH)D concentrations across seasons.
Although it is well established that 25(OH)D concentrations below 20 ng/mL are inadequate and that 25(OH)D concentrations <10 ng/mL are deficient, there is currently no consensus on optimal concentrations of 25(OH)D (18). Concentrations of 20–30 ng/mL are necessary to maximize intestinal calcium absorption and minimize perturbations in parathyroid hormone, calcium, and phosphorus homeostasis (18). Concentrations between 30 and 40 ng/mL have been associated with lower risk of fracture (38,39) and falls (1,40,41).
The primary source of dietary vitamin D in this population was from multivitamins. Although evidence suggests that intakes of at least 800–1000 IU may be needed to obtain 25(OH)D concentrations of 30 ng/mL, approximately 50% of our population had intakes below 400 IU and, without multivitamin use, fewer than 8% of the population had intakes above 600 IU. Among those using multivitamins, approximately 6–12% of the non-black and black elders, respectively, still had 25(OH)D concentrations below 10 ng/mL, relative to fewer than 1.5–3% of those specifically using vitamin D supplements. Although important in both groups, vitamin D inadequacy was more prevalent in the black than the non-black elders in our sample. These black elders were also less likely than the non-black elders to use vitamin supplements (42). These results underscore that vitamin D insufficiency is a significant problem in elders, a growing demographic where prevention of disability is key to reducing hospitalization and nursing home admission.
This study in Boston elders receiving home care is the largest we are aware of that evaluates associations between 25(OH)D concentrations and comprehensive neuropsychological testing. Our findings are consistent with the limited but growing evidence of a relationship between vitamin D and cognitive function that is mechanistically supported. Further study to explore the physical and neuroprotective effects of vitamin D may provide important clues to approaches to maintain independence for community dwelling elderly individuals.
The authors would like to thank the efforts of John Griffith, PhD, Lori Lyn Price, MS, and the Tufts-NEMC General Clinical Research Facility for their support and guidance in the data management of this study and Dr Wendy Qiu for her dedication to the conduct and oversight of the study procedures. J.E.B. designed the analysis and drafted the manuscript; T.M.S. advised on experimental design and interpretation of the cognitive data; G.E.D. advised on data analysis; B.D.H., I.H.R., and M.F.F. provided advice on the content; and K.L.T. oversaw the data analysis and final preparation of the manuscript. All authors reviewed the final version. This research was supported in part by National Institutes of Health grant, AG21790-01; the Tufts New England Medical Center General Clinical Research Center, funded by the National Centerfor Research Resources of the NIH; and by United States Department of Agriculture Agricultural Research Service agreement # 58-1950-7-707.