Alzheimer’s & Dementia 10 (2014) 296–302

Reduced 25-hydroxyvitamin D and risk of Alzheimer’s diseaseand vascular dementia

Shoaib Afzala, Stig E. Bojesena,b,c, Børge G. Nordestgaarda,b,c,*aThe Department of Clinical Biochemistry, Herlev Hospital, Copenhagen University Hospital, Herlev, Denmark

bThe Copenhagen City Heart Study, Bispebjerg Hospital, Copenhagen University Hospital, Copenhagen, DenmarkcFaculty of Health and Medical Sciences, University of Copenhagen, Herlev, Denmark

Abstract Background: Vitamin D deficiency has been implicated as a risk factor for dementia in several

*Corresponding au

E-mail address: B

1552-5260/$ - see fro

http://dx.doi.org/10.10

cross-sectional studies. We tested the hypothesis that reduced plasma 25-hydroxyvitamin D (25[OH]D) is associated with increased risk of Alzheimer’s disease (AD) and vascular dementia inthe general population.Methods: Wemeasured baseline plasma 25(OH)D in 10,186 white individuals from the Danish gen-eral population.Results: During 30 years of follow-up, 418 participants developed AD and 92 developed vascular de-mentia. Multivariable adjusted hazard ratios for AD were 1.25 (95% confidence interval [CI], 0.95–1.64) for 25(OH)D less than 25 nmol/L vs. greater than or equal to 50 nmol/L, and 1.29 (95% CI,1.01–1.66) for less than the 25th seasonally adjusted 25(OH)D percentile vs. more than the 50th sea-sonally adjusted 25(OH)D percentile. Multivariable adjusted hazard ratios for vascular dementiawere1.22 (95%CI, 0.77–1.91) for 25(OH)D less than 50 nmol/L vs. greater than or equal to 50 nmol/L, and1.22 (95% CI, 0.79–1.87) for less than or equal to the 50th vs. more than the 50th seasonally adjusted25(OH)D percentile. Last, multivariable adjusted hazard ratios for the combined end point were 1.28(95%CI, 1.00–1.64) for 25(OH)D less than 25 nmol/L vs. greater than or equal to 50 nmol/L, and 1.27(95% CI, 1.01–1.60) for less than the 25th vs. more than the 50th seasonally adjusted 25(OH)D.Conclusions: We observed an association of reduced plasma 25(OH)D with increased risk of thecombined end point of AD and vascular dementia in this prospective cohort study of the general pop-ulation.� 2014 The Alzheimer’s Association. All rights reserved.

Keywords: Alzheimer’s disease; Vascular dementia; Vitamin D; General population; Cohort study

1. Introduction

The pathogenetic hallmarks of Alzheimer’s disease (AD)are thought to be accumulation of dysfunctional proteins(i.e. amyloid beta [Ab] and tau protein derivates) in the brainfollowed by oxidative damage and inflammation, leading toderanged energy metabolism, localized synaptic failure, andneuronal loss [1]. The hormonally active form of vitaminD—1,25-dihydroxyvitamin D (1,25-(OH)2-vitD)—hasbeen shown to induce Ab removal [2,3], to reduceoxidative stress-induced cell damage present in AD [4,5],

thor. Tel.: 145 38683297; Fax: 145 38683311.

oerge.Nordestgaard@regionh.dk

nt matter � 2014 The Alzheimer’s Association. All rights r

16/j.jalz.2013.05.1765

to improve intracellular Ca21 homeostasis dysregulated inAD [6–8], to upregulate neurotrophic factors [6,9,10], andto induce neuroprotective processes [11–13]. Furthermore,several cross-sectional studies have shown reduced plasmalevels of 25-OH-vitamin D (25[OH]D) among AD patientscompared with control subjects [14–16].

Vascular dementia develops as a result of infarcts [17],and clinical studies have shown that vitamin D deficiencyis associated with white matter hyperintensities in patientswith dementia and risk of stroke [18,19]. Moreover,vitamin D deficiency has been associated with commonrisk factors for stroke—namely, hypertension and diabetes[20,21]—further supporting a possible role for vitamin Ddeficiency in the pathogenesis of vascular dementia.However, this has never been tested previously.

eserved.

S. Afzal et al. / Alzheimer’s & Dementia 10 (2014) 296–302 297

Therefore, we tested the hypothesis that decreasedplasma 25(OH)D is associated with increased risk of ADand vascular dementia in the general population. For thispurpose, we studied 10,186 white individuals from the Co-penhagen City Heart Study monitored for up to 30 yearswithout losses to follow-up.

2. Methods

2.1. Study design

The Copenhagen City Heart Study is a prospective cohortstudy of the Danish general population initiated in 1976 to1978 with follow-up examinations in 1981 to 1983, 1991to 1993, and 2001 to 2003 [22–24]. Individuals 20 to 100years of age were drawn randomly from the nationalDanish Central Person Register and invited to participate;all inhabitants in Denmark are uniquely identified throughtheir central person registration number, which also holdsinformation on age and sex.

The current study included 10,186 participants from the1981 to 1983 examination, who were free of any type of di-agnosed dementia at baseline and had available plasma sam-ples for 25(OH)D measurement. A Danish ethics committeeapproved the study (KF100.2039/91 and KF01-144/01). Par-ticipants provided written informed consent.

2.2. Measurements of 25(OH)D and other biochemicalanalytes

Plasma samples collected at baseline in 1981 to 1983were stored at –20�C until 2009 to 2010, when 25(OH)Dwas measured using the DiaSorin Liaison 25(OH)D TOTALassay (Diasorin, Stillwater, MN). Assay precision wastested daily whereas assay accuracy was tested monthly us-ing an external quality control program. The interassay coef-ficient of variance was 10% for low-level control subjects(w16 nmol/L) and 8% for high-level control subjects(w54 nmol/L).

Colorimetric assays (Boehringer Mannheim, Mannheim,Germany or Konelab, Espoo, Finland) were used to measurecreatinine, total cholesterol, and high-density lipoprotein(HDL) cholesterol in plasma.Total cholesterol andHDLcho-lesterol were measured the same day as the blood samplewascollected, and creatinine was measured during 2009 to 2010using samples stored at 220�C without previous thawing.

2.3. Covariates

Information on smoking habits was obtained from self-reported questionnaires completed together with an examineron the day of attendance. Participants were asked about dura-tion and intensity of leisure time and work-related physicalactivities (hours per week), level of income (high, medium,or low), and level of education (years). Information on base-line diabetesmellitus and hypertensionwas assessed by ques-tions regarding current morbidities, current medications, and

baseline measurements of nonfasting glucose (.11 mmol/L)and blood pressure (.140/90 mmHg or.135/85 mmHg forparticipants with diabetes mellitus). Participants were alsoasked about type and amount ofweekly alcohol consumption,which was then calculated in units of 12 g/week. Body massindexwas calculated asmeasuredweight (kilograms) dividedby measured height (meters) squared.

2.4. End points

Information on incident diagnoses of AD (InternationalClassification of Diseases [ICD], 8th edition, codes 290.0–290.1 and ICD, 10th edition, codes F00 andG30) andvasculardementia (ICD, 8th edition, codes 293.09–293.1 and ICD,10th edition, codes F01) was collected from diagnoses en-tered in the national Danish Patient Registry and the nationalDanish Causes of Death Registry [25]. Validation studieshave shown high validity of the dementia diagnosis and ADdiagnoses in Danish registries [25]. Follow-up time for eachsubject began at the day of blood sampling in 1981 to 1983and ended at diagnosis of AD (n5 418) or vascular dementia(n5 92), death (n5 6360), emigration (n5 55), orMay 2011(n5 3275), whichever occurred first. The median follow-uptime was 21 years (range, 0.03–30 years). Follow-up was100%complete—that is, we did not lose track of even a singleindividual. We also combined AD and vascular dementia forthree reasons: (i) to increase statistical power, (ii)ADand vas-cular dementia show clinical overlap, and (iii) vitamin Dmaybe involved in the pathogenesis of both diseases.

Participants with both diagnoses were included in bothend points separately for two reasons: (i) it is well knownthat some patients have dementia with phenotypes interme-diate between AD and vascular dementia and (ii) excludingthese patients or only including them in one of the groups didnot change the effect estimates from the regression analyses(data not shown).

2.5. Statistical analyses

We divided baseline 25(OH)D into the following a prioriseasonally unadjusted clinical categories of more than50nmol/L (sufficient levels), 25 to 49.9 nmol/L (insufficientlevels), and less than 25 nmol/L (deficient levels). In addition,because concentrations of 25(OH)D vary according to thetime of year because of the high-latitude geographic positionof Denmark, we also used seasonally adjusted 25(OH)Dlevels. Two strategies were applied to adjust for the seasonalvariation in vitamin D. First, we used unadjusted 25(OH)Dlevels in regression analyses and adjusted for calendar monthof blood draw. Second, calendar month-specific cut pointswere obtained by assigning subjects to more than the 50thpercentile, the 26th to the 50th percentile, and less than orequal to the 25th percentile categories within the samemonthof sample collection. For vascular dementia, because of therelatively few cases, we changed classifications to less than50 nmol/L vs. greater than or equal to 50 nmol/L for the

Table 1

Baseline characteristics according to clinical cut points for 25-hydroxyvitamin D plasma levels unadjusted for seasonal variation

Characteristic

Plasma 25-hydroxyvitamin D, nmol/L

,25 (n 5 2384) 25–49.9 (n 5 4087) .50 (n 5 3715) Trend, P value*

Men, % 45 44 43 .10

Age, years 58 (50–65) 58 (49–65) 57 (47–64) ,.001

Ever smoker, % 84 78 77 ,.001

Body mass index, kg/m2 25 (23–29) 25 (23–28) 24 (22–27) ,.001

High physical activity (leisure), % 27 34 40 ,.001

High physical activity (work), % 26 25 25 .46

High income, % 16 21 26 ,.001

Education, years 7 (7–9) 8 (7–10) 8 (7–10) ,.001

Diabetes mellitus, % 5 4 2 ,.001

Hypertension, % 60 56 49 ,.001

Alcohol consumption, units/weeky 3 (0–12) 4 (1–11) 5 (2–11) ,.001

Cholesterol, mmol/L 5.8 (5.1–6.5) 5.9 (5.2–6.7) 5.9 (5.1–6.7) .001

HDL cholesterol, mmol/L 1.1 (0.9–1.3) 1.1 (0.9–1.3) 1.2 (1.0–1.4) ,.001

Creatinine, mmol/L 94 (85–105) 94 (85–104) 93 (85–102) .001

Abbreviation: HDL, high-density lipoprotein.

NOTE. International System of Units conversion factors: To convert 25(OH)D to nanomoles per liter from nanograms per liter, multiply values by 2.5.

Continuous values are shown as median (interquartile range).

*P values were calculated using Cuzick’s nonparametric trend test.yOne unit w 12 g.

S. Afzal et al. / Alzheimer’s & Dementia 10 (2014) 296–302298

clinical categories, and less than or equal to the 50th percen-tile vs. more than the 50th percentile for the seasonally ad-justed percentiles. For trend tests, individuals in each groupwere assigned themedianvalue of their group, either as abso-lute values or as percentiles.

Cox proportional hazards regression was used to estimatehazard ratios with a 95% confidence interval (CI) for incidentAD and vascular dementia. We used age as a timescale withdelayed entry (left truncation). Thus, age differenceswere ad-justed automatically for and referred to in text, tables, and fig-ures as age adjusted.However, for the test of interaction of agewith 25(OH)D levels on risk of dementia, we used years offollow-up as the timescale. Multivariable adjusted Cox re-gression models included gender, age, smoking status(never/ever), body mass index, duration and intensity of lei-sure time andwork-related physical activities, income, educa-tion, diabetes mellitus, hypertension, alcohol consumption,cholesterol, HDL cholesterol, creatinine, and calendar monthof blood draw because these are suspected risk factors for de-mentia and vitamin D deficiency. Interactions were tested us-ing the Wald test with Cox regression models includingmultiplicative two-factor interaction terms. For interactionanalyses and stratified analyses, we used a 20th percentile de-crease in seasonally adjusted percentiles of plasma 25(OH)Das the independent variable. The proportional hazards as-sumption was tested for with Cox regression models usingSchoenfeld residuals; no departures were detected for the dif-ferent plasma 25(OH)D variables used. We analyzed the datawith STATA 12.1 (STATA, College Station, TX).

3. Results

Table 1 summarizes the baseline characteristics by25(OH)D levels. Reduced levels of 25(OH)D were associ-

ated with increasing age, smoking, increasing body mass in-dex, low duration and intensity of leisure time physicalactivity, low income, low education, diabetes mellitus, hy-pertension, reduced alcohol consumption, reduced choles-terol, reduced HDL cholesterol, and increased creatininelevels. A total of 418 events of AD and 92 events of vasculardementia occurred during 30 years of follow-up. Both diag-noses were registered for 14 participants.

Median levels of 25(OH)D were 41 nmol/L among allparticipants and 39 nmol/L among those who developedAD or vascular dementia.

3.1. Alzheimer’s disease

Adjusted hazard ratios for AD increased with decreasinglevels of 25(OH)D by clinical categories and by seasonallyadjusted percentile categories (Tables 2 and 3), althoughonly with a significant trend when using seasonally adjustedpercentile categories. Multivariable adjusted hazard ratioswere 1.25 (95% CI, 0.95–1.64) for 25(OH)D less than 25nmol/L vs. greater than or equal to 50 nmol/L, and 1.29(1.01–1.66) for less than the 25th vs. more than the 50th sea-sonally adjusted percentile in the fully adjusted models.Multivariable adjusted hazard ratios were 1.04 (1.00–1.09)per 10-nmol/L decrease in 25(OH)D and 1.08 (1.01–1.16)per 20th-percentile decrease in seasonally adjusted percen-tiles of 25(OH)D in the fully adjusted models.

3.2. Vascular dementia

Multivariable adjusted hazard ratios for vascular demen-tia were 1.22 (95% CI, 0.77–1.91) for 25(OH)D less than 50nmol/L vs. greater than or equal to 50 nmol/L, and 1.22(95% CI, 0.79–1.87 for less than or equal to the 50th vs.

Table 2

Risk of Alzheimer’s disease, vascular dementia, and the combined end point by plasma 25-hydroxyvitamin D (25[OH]D) in clinical categories unadjusted for

seasonal variation

End point

25(OH)D,

nmol/L Participants, n Events, n

Model 1,* age and sex adjusted Model 2,y partially adjusted Model 3,z fully adjusted

HR (95% CI) P valuex HR (95% CI) P valuex HR (95% CI) P valuex

Alzheimer’s

disease

�50 3715 151 1 .12 1 .12 1 .11

25–49.9 4087 174 1.11 (0.89–1.39) 1.11 (0.89–1.39) 1.12 (0.90–1.40)

,25 2384 93 1.23 (0.95–1.59) 1.23 (0.94–1.61) 1.25 (0.95–1.64)

Vascular

dementia

�50 3715 31 1 .25 1 .44 1 .45

,50 6410 61 1.29 (0.83–1.99) 1.19 (0.76–1.86) 1.22 (0.77–1.91)

Combined �50 3715 175 1 .04 1 .06 1 .04

25–49.9 4087 210 1.16 (0.95–1.42) 1.15 (0.93–1.41) 1.16 (0.94–1.42)

,25 2384 111 1.26 (0.99–1.61) 1.25 (0.98–1.60) 1.28 (1.00–1.64)

Abbreviations: HR, hazard ratio; CI, confidence interval.

NOTE. International System of Units conversion factors: To convert 25(OH)D to nanomoles per liter from nanograms per liter, multiply values by 2.5.

*Adjusted for age, sex, and month of blood sample only.yModel 2 adjustment: model 1 covariates 1 smoking status, body mass index, leisure time and work-related physical activity, and alcohol consumption.zModel 3 adjustment: model 2 covariates 1 income level, education, baseline diabetes mellitus, hypertension, cholesterol, high-density lipoprotein choles-

terol, and creatinine.xP values are from trend tests.

S. Afzal et al. / Alzheimer’s & Dementia 10 (2014) 296–302 299

more than the 50th seasonally adjusted percentile (Tables 2and 3). Multivariable adjusted hazard ratios were 1.02 (95%CI, 0.92–1.12) per 10-nmol/L decrease in 25(OH)D and 1.05(95% CI, 0.90–1.23) per 20th-percentile decrease in season-ally adjusted percentiles of 25(OH)D in the fully adjustedmodels.

3.3. Combined end point

Adjusted hazard ratios for AD or vascular dementia (i.e.,the combined end point) increased with decreasing levels of25(OH)D by clinical categories and by seasonally adjustedpercentile categories (Tables 2 and 3). Multivariable ad-justed hazard ratios were 1.28 (95% CI, 1.00–1.64) for

Table 3

Risk of Alzheimer’s disease, vascular dementia, and the combined end point by p

categories

End point 25(OH)D, % Participants, n Events, n

Model 1,* age and sex

HR (95% CI)

Alzheimer’s

disease

.50 5093 132 1

26–50 2545 151 1.17 (0.93–1.48)

�25 2548 135 1.23 (0.97–1.56)

Vascular

dementia

.50 5093 44 1

�50 5093 48 1.26 (0.83–1.90)

Combined .50 5093 156 1

26–50 2545 177 1.20 (0.97–1.48)

�25 2548 163 1.23 (0.98–1.53)

Abbreviations: HR, hazard ratio; CI, confidence interval.

*Adjusted for age and sex only.yModel 2 adjustment: model 1 covariates 1 smoking status, body mass index,zModel 3 adjustment: model 2 covariates 1 income level, education, baseline d

terol, and creatinine.xP values are from trend tests.

25(OH)D less than 25 nmol/L vs. greater than or equal to50 nmol/L, and 1.27 (95% CI, 1.01–1.60) for less than the25th vs. more than the 50th seasonally adjusted percentilein the fully adjusted models. Multivariable adjusted hazardratios were 1.04 (95% CI, 1.00–1.09) per 10-nmol/L de-crease in 25(OH)D and 1.08 (95% CI, 1.01–1.15) per20th-percentile decrease in seasonally adjusted percentilesof 25(OH)D in the fully adjusted models.

In all stratified analyses, a 20th-percentile reduction inseasonally adjusted percentile of plasma 25(OH)D was asso-ciated with hazard ratios nominally above 1.00 except forparticipants with diabetes mellitus (Fig. 1). There were nointeractions of 25(OH)D with other variables on risk of thecombined end point.

lasma 25-hydroxyvitamin D (25[OH]D) in seasonally adjusted percentile

adjusted Model 2,y partially adjusted Model 3,z fully adjusted

P valuex HR (95% CI) P valuex HR (95% CI) P valuex

.07 1 .04 1 .03

1.20 (0.95–1.51) 1.23 (0.97–1.55)

1.25 (0.98–1.60) 1.29 (1.01–1.66)

.26 1 .42 1 .42

1.19 (0.78–1.81) 1.22 (0.79–1.87)

.04 1 .04 1 .02

1.21 (0.98–1.50) 1.24 (1.00–1.54)

1.23 (0.98–1.54) 1.27 (1.01–1.60)

leisure time and work-related physical activity, and alcohol consumption.

iabetes mellitus, hypertension, cholesterol, high-density lipoprotein choles-

Fig. 1. Risk of Alzheimer’s disease or vascular dementia (combined end point) by plasma 25-hydroxyvitamin D (25[OH]D) in strata. Each hazard ratio is per

20th percentile decrease in seasonally adjusted plasma 25(OH)D percentiles adjusted for gender, age, smoking status (never/ever), body mass index (BMI),

duration and intensity of leisure time and work-related physical activities, income, education, diabetes mellitus, hypertension, alcohol consumption, cholesterol,

high-density lipoprotein (HDL) cholesterol, and creatinine, excluding the variable used for stratification. Age, education, alcohol consumption, cholesterol, and

HDL cholesterol was categorized using the median. The size of each square corresponds to the size of the population compared with the total population—based

on 10,186 individuals from the Danish general population in the Copenhagen City Heart Study monitored for up to 30 years after blood sampling for measure-

ment of 25(OH)D. CI, confidence interval.

S. Afzal et al. / Alzheimer’s & Dementia 10 (2014) 296–302300

4. Discussion

In a prospective study on the risk of AD and vasculardementia in the general population with reduced levels of25(OH)D, we observed an increasing risk of AD withdecreasing levels of 25(OH)D.

Biologically, our results are plausible for AD because in-creased 1,25-(OH)2-vitD has been implicated in increasedclearance of Ab by macrophages [2,3]; inhibition ofmechanisms of free-radical production induced in, for exam-ple, AD [4,5]; upregulation of synaptic neurotrophic factors,which are depleted in AD [6,9,10]; and, last, protectionof neurons from apoptosis induced by, for example, Ab[11–13]. For vascular dementia, clinical studies haveshown an association of reduced plasma 25(OH)D withincreased risk of stroke and white matter hyperintensities[18,19], and in vivo studies have shown that higher plasma25(OH)D levels may restrict the size of experimentallyinduced infarctions [26]. Furthermore, plasma 25(OH)D orvitamin D intake has been shown to be associated with cog-nitive function [27–31] and cognitive decline [32–34]

among the elderly population, supporting indirectly a rolefor plasma 25(OH)D in maintenance of cognitive function.

Our results are in accordance with most previous cross-sectional studies showing reduced levels of 25(OH)D inpatients with AD or cognitive impairment [14–16].Furthermore, studies have shown an association of lowintake of vitamin D with increased risk of AD [31,35], andprospective studies have shown an association of lowplasma 25(OH)D with cognitive decline [34,36]. Also, thecurrent prospective association of reduced levels ofvitamin D with vascular dementia did not reach statisticalsignificance, probably because of low power. Previousstudies showed an association of reduced levels of 25(OH)D with stroke [37], which is a prerequisite for developingvascular dementia [17].

A major limitation of this study was relying on cliniciandiagnoses of dementia for our primary outcomes. This is par-ticularly problematic when relying on clinicians to make anadequate distinction between AD and vascular dementiawithout any verification of imaging studies, and the low

S. Afzal et al. / Alzheimer’s & Dementia 10 (2014) 296–302 301

number of vascular dementia cases seems to reflect this.However, validation studies have shown high validity ofthe dementia diagnosis and AD diagnoses in Danish regis-tries [25]. Other potential limitations of our study includehealthy participant bias because participants are typicallyhealthier than background populations; however, this wouldtend to weaken the associations rather than strengthen them,and thus is unlikely to explain our findings. The delay inmeasurement from 1981 to 1983 to 2009 to 2010 could raiseconcern of decay in plasma 25(OH)D, but this seems un-likely for several reasons: we noticed the expected seasonalvariation of 25(OH)D levels [38,39]; median concentrationsof plasma 25(OH)D across plasma samples from threedifferent examinations on the same healthy participantswith storage times of 10, 20, and 30 years were similar[38,39]; previous studies have shown high stability duringstorage [40]; the median concentration observed in our studyof 16 nmol/L was similar to that in comparable populations[41]; and a reduced quality of the 25(OH)D measurementwould tend to weaken rather than inflate an association. An-other potential limitation was the use of only a single mea-surement of plasma 25(OH)D, although studies haveshown that a single measurement is a reasonable predictorof long-term exposure to plasma 25(OH)D [42]. It is alsoa potential limitation that we were not able to adjust for de-pression. Furthermore, all participants were white, limitingthe applicability of our results to other ethnic groups. The di-agnoses were based on registry data and there is a high prob-ability of underdiagnosis of dementia, which could affectour results. However, admixture of cases and noncaseswould tend to weaken the associations rather than strengthenthem and thus is unlikely to explain our findings.

The strengths of our study are that the study was conduct-ed on the general population, we had up to 30 years offollow-up with no losses to follow-up, we could accountfor other major risk factors associated with risk of dementia,and we had high statistical power to examine the prospectiveassociations of reduced plasma 25(OH)D levels with AD.Furthermore, in northern Europe ultraviolet B radiationfrom the sun is only adequate for sufficient endogenous vi-tamin D production in the skin during the summer months(May–September) and food has never been fortified with vi-tamin D in Denmark. Thus, this cohort from the Danish gen-eral population allows determination of the natural history ofthe association of vitamin D deficiency with risk of AD andvascular dementia.

In conclusion, we observed an association of reducedplasma 25(OH)D with increased risk of AD and vascular de-mentia in this prospective cohort study of the general popu-lation. However, these preliminary results requireconfirmation in other studies.

Acknowledgments

TheDanishHeart Foundation andHerlevHospital, Copenha-gen University Hospital provided research support. Diasorin

Liaison provided kits for measurement of 25(OH)D. Thesponsors had no role in the study design, statistical analysis,data interpretation,manuscript drafting,manuscript revision,or the decision to submit the manuscript for publication.

RESEARCH IN CONTEXT

1. Systematic review: We searched PubMed up untilDecember 01, 2012, for articles published in any lan-guagewith the search terms Vitamin D and Dementiaor Alzheimer’s disease or Vascular dementia. Wealso reviewed the reference lists of the documentsidentified by this search.

2. Interpretation: Our prospective study of reduced 25-hydroxyvitamin D and risk of Alzheimer’s disease(AD) and vascular dementia extends epidemiologicknowledge of the association between vitamin Dstatus and specific types of dementia. Our findingssuggest that a low level of 25-hydroxyvitamin D maybe a risk factor for future risk of AD.

3. Future directions: In our study, we had 418 patientswith AD, which is more than most previous studieson this topic. However, given the size of the risk es-timates, larger studies are required, especially re-garding the possible role of 25-hydroxyvitamin Din the development of vascular dementia.

References

[1] Querfurth HW, LaFerla FM. Alzheimer’s disease. N Engl J Med 2010;

362:329–44.

[2] Masoumi A, Goldenson B, Ghirmai S, Avagyan H, Zaghi J, Abel K,

et al. 1alpha,25-dihydroxyvitamin D3 interacts with curcuminoids to

stimulate amyloid-beta clearance by macrophages of Alzheimer’s dis-

ease patients. J Alzheimers Dis 2009;17:703–17.

[3] Mizwicki MT, Menegaz D, Zhang J, Barrientos-Duran A, Tse S,

Cashman JR, et al. Genomic and nongenomic signaling induced by

1alpha,25(OH)2-vitamin D3 promotes the recovery of amyloid-beta

phagocytosis by Alzheimer’s disease macrophages. J Alzheimers Dis

2012;29:51–62.

[4] Garcion E, Nataf S, Berod A, Darcy F, Brachet P. 1,25-Dihydroxyvi-

tamin D3 inhibits the expression of inducible nitric oxide synthase

in rat central nervous system during experimental allergic encephalo-

myelitis. Brain Res Mol Brain Res 1997;45:255–67.

[5] Meda L, Cassatella MA, Szendrei GI, Otvos L Jr, Baron P, Villalba M,

et al. Activation of microglial cells by beta-amyloid protein and inter-

feron-gamma. Nature 1995;374:647–50.

[6] Gezen-Ak D, Dursun E, Yilmazer S. The effects of vitamin D receptor

silencing on the expression of LVSCC-A1C and LVSCC-A1D and the

release of NGF in cortical neurons. PLoS One 2011;6:e17553.

[7] Brewer LD, Porter NM, Kerr DS, Landfield PW, Thibault O. Chronic

1alpha,25-(OH)2 vitamin D3 treatment reduces Ca2+ -mediated hip-

pocampal biomarkers of aging. Cell Calcium 2006;40:277–86.

S. Afzal et al. / Alzheimer’s & Dementia 10 (2014) 296–302302

[8] Brewer LD, Thibault V, Chen KC, Langub MC, Landfield PW,

Porter NM. Vitamin D hormone confers neuroprotection in parallel

with downregulation of L-type calcium channel expression in hippo-

campal neurons. J Neurosci 2001;21:98–108.

[9] Saporito MS, Brown ER, Hartpence KC, Wilcox HM, Vaught JL,

Carswell S. Chronic 1,25-dihydroxyvitamin D3-mediated induction

of nerve growth factor mRNA and protein in L929 fibroblasts and in

adult rat brain. Brain Res 1994;633:189–96.

[10] Musiol IM, Feldman D. 1,25-dihydroxyvitamin D3 induction of nerve

growth factor in L929 mouse fibroblasts: effect of vitamin D receptor

regulation and potency of vitamin D3 analogs. Endocrinology 1997;

138:12–8.

[11] Dursun E, Gezen-Ak D, Yilmazer S. A novel perspective for Alz-

heimer’s disease: vitamin D receptor suppression by amyloid-beta

and preventing the amyloid-beta induced alterations by vitamin D in

cortical neurons. J Alzheimers Dis 2011;23:207–19.

[12] Obradovic D, Gronemeyer H, Lutz B, Rein T. Cross-talk of vitamin D

and glucocorticoids in hippocampal cells. J Neurochem 2006;

96:500–9.

[13] Taniura H, Ito M, Sanada N, Kuramoto N, Ohno Y, Nakamichi N, et al.

Chronic vitamin D3 treatment protects against neurotoxicity by gluta-

mate in association with upregulation of vitaminD receptor mRNA ex-

pression in cultured rat cortical neurons. J Neurosci Res 2006;

83:1179–89.

[14] Annweiler C, Llewellyn DJ, Beauchet O. Low serum vitamin D con-

centrations in Alzheimer’s disease: a systematic review andmeta-anal-

ysis. J Alzheimers Dis 2013;33:659–74.

[15] Balion C, Griffith LE, Strifler L, Henderson M, Patterson C,

Heckman G, et al. Vitamin D, cognition, and dementia: a systematic

review and meta-analysis. Neurology 2012;79:1397–405.

[16] Etgen T, Sander D, Bickel H, Sander K, Forstl H. Vitamin D defi-

ciency, cognitive impairment and dementia: a systematic review and

meta-analysis. Dement Geriatr Cogn Disord 2012;33:297–305.

[17] Vermeer SE, Prins ND, den HT, Hofman A, Koudstaal PJ,

Breteler MM. Silent brain infarcts and the risk of dementia and cogni-

tive decline. N Engl J Med 2003;348:1215–22.

[18] Buell JS, Dawson-Hughes B, Scott TM, Weiner DE, Dallal GE,

Qui WQ, et al. 25-Hydroxyvitamin D, dementia, and cerebrovascular

pathology in elders receiving home services. Neurology 2010;

74:18–26.

[19] Sun Q, Pan A, Hu FB, Manson JE, Rexrode KM. 25-Hydroxyvitamin

D Levels and the Risk of Stroke. Stroke 2012;43:1470–7.

[20] Burgaz A, Orsini N, Larsson SC,Wolk A. Blood 25-hydroxyvitamin D

concentration and hypertension: a meta-analysis. J Hypertens 2011;

29:636–45.

[21] Mitri J, Muraru MD, Pittas AG. Vitamin D and type 2 diabetes: a sys-

tematic review. Eur J Clin Nutr 2011;65:1005–15.

[22] Schnohr P, Jensen JS, Scharling H, Nordestgaard BG. Coronary heart

disease risk factors ranked by importance for the individual and com-

munity. A 21 year follow-up of 12 000 men and women from The Co-

penhagen City Heart Study. Eur Heart J 2002;23:620–6.

[23] Nordestgaard BG, Benn M, Schnohr P, Tybjaerg-Hansen A. Nonfast-

ing triglycerides and risk of myocardial infarction, ischemic heart dis-

ease, and death in men and women. JAMA 2007;298:299–308.

[24] Bojesen SE, Tybjaerg-Hansen A, Nordestgaard BG. Integrin beta3

Leu33Pro homozygosity and risk of cancer. J Natl Cancer Inst 2003;

95:1150–7.

[25] Phung TK, Andersen BB, Hogh P, Kessing LV, Mortensen PB,

Waldemar G. Validity of dementia diagnoses in the Danish hospital

registers. Dement Geriatr Cogn Disord 2007;24:220–8.

[26] Wang Y, Chiang YH, Su TP, Hayashi T, Morales M, Hoffer BJ,

et al. Vitamin D(3) attenuates cortical infarction induced by middle

cerebral arterial ligation in rats. Neuropharmacology 2000;

39:873–80.

[27] Annweiler C, Schott AM, Allali G, Bridenbaugh SA, Kressig RW,

Allain P, et al. Association of vitamin D deficiency with cognitive im-

pairment in older women: cross-sectional study. Neurology 2010;

74:27–32.

[28] Llewellyn DJ, Langa KM, Lang IA. Serum 25-hydroxyvitamin D con-

centration and cognitive impairment. J Geriatr Psychiatry Neurol

2009;22:188–95.

[29] Llewellyn DJ, Lang IA, Langa KM, Melzer D. Vitamin D and cogni-

tive impairment in the elderly U.S. population. J Gerontol A Biol Sci

Med Sci 2011;66:59–65.

[30] Buell JS, Scott TM, Dawson-Hughes B, Dallal GE, Rosenberg IH,

Folstein MF, et al. Vitamin D is associated with cognitive function

in elders receiving home health services. J Gerontol A Biol Sci Med

Sci 2009;64:888–95.

[31] Annweiler C, Schott AM, Rolland Y, Blain H, Herrmann FR,

Beauchet O. Dietary intake of vitamin D and cognition in older

women: a large population-based study. Neurology 2010;

75:1810–6.

[32] Slinin Y, Paudel ML, Taylor BC, Fink HA, Ishani A, CanalesMT, et al.

25-Hydroxyvitamin D levels and cognitive performance and decline in

elderly men. Neurology 2010;74:33–41.

[33] Breitling LP, Perna L, Muller H, Raum E, Kliegel M, Brenner H. Vi-

tamin D and cognitive functioning in the elderly population in Ger-

many. Exp Gerontol 2012;47:122–7.

[34] Llewellyn DJ, Lang IA, Langa KM, Muniz-Terrera G, Phillips CL,

Cherubini A, et al. Vitamin D and risk of cognitive decline in elderly

persons. Arch Intern Med 2010;170:1135–41.

[35] Annweiler C, Rolland Y, Schott AM, Blain H, Vellas B, Herrmann FR,

et al. Higher vitamin D dietary intake is associated with lower risk of

alzheimer’s disease: a 7-year follow-up. J Gerontol A Biol Sci Med Sci

2012;67:1205–11.

[36] Slinin Y, Paudel M, Taylor BC, Ishani A, RossomR, Yaffe K, et al. As-

sociation between serum 25(OH) vitamin D and the risk of cognitive

decline in older women. J Gerontol A Biol Sci Med Sci 2012;

67:1092–8.

[37] Brondum-Jacobsen P, Nordestgaard BG, Schnohr P, Benn M. 25-Hy-

droxyvitamin D and symptomatic ischemic stroke: An Original Study

and Meta-Analysis. Ann Neurol 2013;73:38–47.

[38] Afzal S, Bojesen SE, Nordestgaard BG. Low 25-hydroxyvitamin d and

risk of type 2 diabetes: a prospective cohort study and metaanalysis.

Clin Chem 2013;59:381–91.

[39] Afzal S, Nordestgaard BG, Bojesen SE. Plasma 25-hydroxyvitamin D

and risk of non-melanoma and melanoma skin cancer: a prospective

cohort study. J Invest Dermatol 2013;133:629–36.

[40] OckeMC, Schrijver J. Obermann-de Boer GL, Bloemberg BP, Haenen

GR, Kromhout D. Stability of blood (pro)vitamins during four years of

storage at -20 degrees C: consequences for epidemiologic research. J

Clin Epidemiol 1995;48:1077–85.

[41] McCullough ML, Weinstein SJ, Freedman DM, Helzlsouer K,

Flanders WD, Koenig K, et al. Correlates of circulating 25-hydroxyvi-

tamin D: Cohort Consortium Vitamin D Pooling Project of Rarer Can-

cers. Am J Epidemiol 2010;172:21–35.

[42] Jorde R, SneveM,HutchinsonM, EmausN, FigenschauY, Grimnes G.

Tracking of serum 25-hydroxyvitamin D levels during 14 years in

a population-based study and during 12 months in an intervention

study. Am J Epidemiol 2010;171:903–8.