Cerebral small vessel disease in middle age and genetic

predisposition to late-onset Alzheimer’s Disease

Authors and Affiliations

James D. Stefaniaka,1, Li Sub,c,1, Elijah Makb, Nasim Sheikh Bahaeid, Katie Wellse, Karen

Ritchief,g,h, Adam Waldmanh, Craig W. Ritchieh, John T. O’Brienb,i*

aManchester Academic Health Sciences Centre, Salford Royal NHS Foundation Trust, Salford,

UK ;

bDepartment of Psychiatry, University of Cambridge School of Clinical Medicine, Cambridge,

UK ;

cChina-UK Centre for Cognition and Ageing Research, Faculty of Psychology, Southwest

University, Chongqing, China ;

dDepartment of Radiology, University of Cambridge School of Clinical Medicine, Cambridge,

UK ;

eThe Centre for Mental Health, Imperial College, London, UK ;

fINSERM Unit 1061 Neuropsychiatry, Montpellier, France ;

gUniversity of Montpellier, France ;

hCentre for Dementia Prevention, Centre for Clinical Brain Sciences, University of Edinburgh,

Edinburgh, UK ;

iCambridgeshire and Peterborough NHS Foundation Trust, Cambridge, UK.

1Joint first authors.

1

*Correspondence to: Professor John T O’Brien, Department of Psychiatry, University of

Cambridge School of Clinical Medicine, Box 189, Level E4 Cambridge Biomedical Campus,

Cambridge CB2 0SP, UK. Email address: john.obrien@medschl.cam.ac.uk

Word count (excluding Title Page, Abstract, Table 1, Contributorship statement,

Acknowledgements, References, Footnotes, Figure Legend, Table Legend): 1449

Abstract word count: 145

Number of figures: 1

Number of tables: 1

Number of references: 16

Number of supplementary files: 1

2

ABSTRACT

INTRODUCTION: Cerebral small vessel disease (CSVD) is associated with late-onset

Alzheimer’s Disease (LOAD), and might contribute to the relationship between APOEє4 and

LOAD, in older people. However, it is unclear whether CSVD begins in middle age in

individuals genetically predisposed to LOAD.

METHODS: We assessed the relationship between radiological markers of CSVD, white

matter hyperintensities and microbleeds, and genetic predisposition to LOAD in a cross-

sectional analysis of cognitively normal subjects aged 40-59 recruited from the PREVENT

Dementia study.

RESULTS: Microbleed prevalence was 14.5% and mean±SD white matter hyperintensity

percentage of total brain volume was 0.41+0.28%. There was no significant association

between APOEє4 carrier status or history of parental dementia and white matter

hyperintensity volume (p=0.713, 0.912 respectively) or microbleeds (p=0.082, 0.562

respectively) on multiple regression.

DISCUSSION: Genetic predisposition to LOAD, through APOE genotype or AD family history,

is not associated with CSVD in middle age.

Keywords: Dementia; White matter hyperintensity; Cerebral microbleed; MRI; Cerebral

small vessel disease; Middle age; Risk factors

3

1. Introduction

Cerebral small vessel disease (CSVD) is associated with late-onset Alzheimer’s Disease

(LOAD) in older people [1] and might contribute to the relationship between APOEє4 and

LOAD [2]. Furthermore, dominantly inherited Alzheimer’s Disease is associated with

regionally increased white matter hyperintensity burden decades before symptom onset

when cognition is normal [3], raising the possibility that CSVD might be an early feature in

the pathogenesis of Alzheimer’s Disease (AD). However, it is unknown whether CSVD

similarly begins in middle age in individuals genetically predisposed to LOAD

We assessed the relationship between key markers of CSVD seen on 3T Magnetic Resonance

Imaging (MRI) (white matter hyperintensities and cerebral microbleeds [CMBs]) and genetic

predisposition to LOAD in cognitively normal subjects aged 40-59 recruited from the

PREVENT Dementia study [4].

Unexpectedly, we found that CSVD is not associated with the main genetic predisposition to

LOAD (APOEє4 carrier status) or dementia family history in cognitively normal middle-aged

subjects.

2. Methods

2.1 Setting and participants

Data was available from PREVENT Dementia subjects; full details of the study are described

elsewhere [4], but participants are cognitively normal, middle-aged (40-59 years) subjects

with or without parental Alzheimer’s/mixed dementia. 160 participants were included for

white matter hyperintensity analysis and 157 (of the same 160 participants) for CMB

4

analysis. The research was approved by the London-Camberwell St Giles NHS ethics

committee. All subjects provided written informed consent.

2.2 MRI acquisition

All subjects were scanned on a 3T Siemens-Verio scanner. MRI parameters are described in

the Supplementary File. All MRIs were reported by a neuroradiologist and analysed using

ITK-SNAP software in random order by a single rater blinded to all study data including

clinical and genetic information.

2.3 White matter hyperintensity analysis

White matter hyperintensity volumes were quantified using Statistical Parametric Mapping

(SPM8) on Fluid-Attenuated Inversion Recovery (FLAIR) MRIs using an automated, validated

method [5,6]. Following brain segmentation white matter hyperintensity volumes were

calculated by applying an intensity threshold of 1.2x the modal intensity. FLAIR

segmentations were manually checked for errors and corrected (Figure 1a-c).

2.4 CMB analysis

The number of CMBs at each topographical location were rated using the Microbleed

Anatomical Rating Scale (MARS) [7] on Susceptibility Weighted Imaging (SWI) MRIs (Figure

1d-e). To increase accuracy, only definite CMBs were included in statistical analyses.

To examine the accuracy of CMB ratings, 40 participants (including those with and without

CMBs) were independently rated by a second reader (a neuroradiologist) blinded to clinical

information. The intraclass correlation coefficient (ICC) for definite CMBs was 0.95,

5

indicating excellent inter-rater reliability. To examine intra-rater reliability, each participant

was rated twice at 2 weeks apart, yielding an ICC of 0.98.

2.5 Statistical analysis

SPSS23 was used. Statistical significance was defined as p<0.05 with Bonferroni correction.

To answer whether white matter hyperintensity burden (defined as “total white matter

hyperintensity volume as a percentage of total brain volume”, WMH) was associated with

genetic predisposition to LOAD, we performed bivariate analyses against WMH using Mann-

Whitney U test for categorical participant characteristics (gender, APOEє4 carrier status,

APOEє2 carrier status, parental AD/mixed dementia) and Spearman rank correlation

coefficient for continuous participant characteristics (age, number of years of education).

These six participant characteristics were then included in multiple linear regression against

log10WMH.

To answer whether CMB burden (defined as “presence of at least one definite CMB”, or

“definite CMB presence”) was associated with genetic predisposition to LOAD, we

performed bivariate analyses against definite CMB presence using Pearson’s Chi-Square test

(or, if insufficient sample size, Fisher’s exact test) for categorical participant characteristics

and binomial logistic regression for continuous participant characteristics. These six

participant characteristics were then included in multiple logistic regression against definite

CMB presence.

(See online supplement for additional methodological details).

3. Results

6

On bivariate analysis (Mann-Whitney U test), median WMH did not differ significantly

(U=1506, z=-0.284, p=0.776) between patients with (0.39%) and without (0.39%) definite

CMBs.

Mean (SD) WMH was 0.41% (0.28) in the study cohort. The only participant characteristic

associated with WMH on bivariate analysis (Table 1a) was age (rs=0.216, p=0.006); gender

(p=0.034, WMH greater in male subjects) did not survive Bonferroni correction (p<0.008).

Similarly, in multiple linear regression against log10WMH (Table 1a), only age (t=2.426,

p=0.016) added significantly to the model; APOEє4 carrier status (t=0.369, p=0.713) and

parental AD/mixed dementia (t=-0.111, p=0.912) did not.

The prevalence of CMB in the study cohort was 14.5%. This is comparable to the few

existing studies which have measured CMB presence in middle aged participants [8]. No

participant characteristic was associated with definite CMB presence on bivariate analysis

(Table 1b); the association with age at baseline (W=6.590, p=0.010) did not survive

Bonferroni correction (p<0.008). In multiple logistic regression against definite CMB

presence (Table 1b) only age at baseline was significantly (W=5.800, p=0.016) associated.

APOEє4 carrier status (W=3.030, p=0.082) and parental AD/mixed dementia (W=0.336,

p=0.562) were not significantly associated with definite CMB presence.

7

Table 1

1a. Comparisons of participant characteristics with WMH

Participant characteristic Number of

participants

Bivariate analyses with WMH Multiple linear regression against log10WMH

WMH (%) (median, IQR) P value T value P value

Age at baseline 160 0.389 (0.244) 0.006 2.426 0.016

Number of years of education 160 0.389 (0.244) 0.793 0.947 0.345

Gender: 0.034* -1.675 0.096

Male 49 0.409 (0.232)

Female 111 0.337 (0.237)

APOEє2 carrier status: 0.082 1.251 0.213

Non-carrier 143 0.382 (0.223)

Carrier 17 0.434 (0.205)

APOEє4 carrier status: 0.580 0.369 0.713

Non-carrier 100 0.396 (0.255)

Carrier 60 0.372 (0.209)

Parental AD/mixed dementia: 0.740 -0.111 0.912

8

No parent with AD/mixed dementia 92 0.392 (0.269)

At least one parent with AD/mixed dementia 68 0.383 (0.219)

1b. Comparisons of participant characteristics with definite CMB presence

Participant characteristic Number of

participants

Bivariate analyses with definite CMB

presence

Multiple logistic regression against definite CMB

presence

No definite

CMB

Definite CMB

present

P value W value P value Odds ratio (95% CI)

Age at baseline in years (mean, SD) 157 51.2 (5.6) 54.6 (5.1) 0.010* 5.800 0.016 1.14 (1.02-1.26)

Number of years of education (mean, SD) 157 16.1 (3.3) 14.9 (4.0) 0.119 1.424 0.233 0.92 (0.81-1.05)

Gender: 0.155

Male (number of participants, %) 49 39 (79.6) 10 (20.4) Reference Reference Reference

Female (number of participants, %) 108 97 (88.2) 13 (11.8) 1.111 0.292 0.60 (0.23-1.56)

APOEє2 carrier status: 0.280

Non-carrier (number of participants, %) 140 121 (86.4) 19 (13.6) Reference Reference Reference

Carrier (number of participants, %) 17 13 (76.5) 4 (23.5) 0.557 0.456 1.65 (0.44-6.12)

APOEє4 carrier status: 0.272

9

Non-carrier (number of participants, %) 98 86 (87.8) 12 (12.2) Reference Reference Reference

Carrier (number of participants, %) 59 48 (81.4) 11 (18.6) 3.030 0.082 2.39 (0.90-6.35)

Parental AD/mixed dementia: 0.609

No parent with AD/mixed dementia (number

of participants, %)

89 75 (84.3) 14 (15.7) Reference Reference Reference

At least one parent with AD/mixed dementia

(number of participants, %)

68 61 (87.1) 9 (12.9) 0.336 0.562 0.75 (0.29-1.97)

*Not statistically significant after Bonferroni correction for multiple comparisons (p<0.008)

10

4. Discussion

We found no significant relationship between markers of CSVD (either CMB or WMH) and

either APOE status or dementia family history in cognitively normal, middle-aged

individuals. This is a novel finding; no previous study has investigated the association

between CMB or WMH and APOE genotype or AD family history in a cohort of cognitively

normal, dementia- and stroke-free participants as young as ours (mean age 51.6 years). The

‘next youngest’ study investigating this association in cognitively normal subjects [9] had a

mean participant age of 58 and found that WMH distribution was associated with APOEє2;

CMBs were not assessed. Similarly, a meta-analysis of 42 studies, including both cognitively

normal and impaired elderly participants, found that APOEє4 carrier status is associated

with increasing WMH and CMB burden [2]. The present study therefore suggests that APOE

genotype does not become associated with CSVD until closer to the age at which LOAD

typically develops.

This finding is relevant to the researcher because it adds to our understanding of the

complex relationship between APOE genotype, amyloid deposition, blood-brain barrier

(BBB) dysfunction and CSVD. Previous studies have suggested that impaired amyloid

clearance along perivascular drainage pathways, as might hypothetically occur in CSVD or

with BBB dysfunction, can accelerate amyloid accumulation [10]. In keeping with this,

Grimmer et al [11] found that WMH volume was associated with the rate of amyloid

deposition in AD. Severe cerebral amyloid angiopathy (CAA) at post-mortem is associated

with increased CSVD even within the same APOE genotype [12], suggesting an APOE-

independent mechanism by which CAA might lead to CSVD. Similar conclusions were drawn

following in vivo Positron Emission Tomography (PET)/MRI in patients with CAA [13]. There

11

is also evidence against an association between APOE genotype and BBB dysfunction in

patients with AD or CSVD [14]. However, Kester et al [15] found that both WMH and CMB

presence is associated with lower cerebrospinal fluid (CSF) amyloid-β 42 in cognitively

normal APOEє4 carriers but not in non-carriers, thus suggesting ‘pathophysiological synergy’

[15] between APOE genotype, CSVD and amyloid deposition that exists even in the absence

of cognitive impairment. Incorporation of amyloid imaging and BBB integrity data in future

analyses of the present study would lend further insights into the complex relationship

between these variables.

Increased total WMH volume can be detected 6 years, and increased posterior WMH

volume can be detected 22 years, before estimated age of onset of symptom onset in

dominantly inherited AD [3], suggesting that CSVD is an early feature of dominantly

inherited AD. Furthermore, we know that APOEє4 carriers have statistically significantly

smaller hippocampal volumes as early as the third decade of life [16], suggesting that

pathogenic mechanisms predisposing to LOAD in APOEє4 carriers will have started many

years before our middle-aged participants were enrolled in PREVENT Dementia. Our finding

that CSVD is not associated with APOE genotype in middle age therefore suggests that CSVD

is not present during the early stages of the pathogenic mechanisms predisposing to LOAD

in APOEє4 carriers. This is of interest to the clinician considering therapeutic targets for the

prevention of LOAD.

An important limitation to our finding is the cross-sectional nature of our data meaning that

we do not know whether individuals with genetic predisposition to LOAD, as measured by

APOE genotype and AD family history, will actually go on to develop LOAD. The sample size

is restricted and these findings need to be replicated. Therefore the PREVENT Dementia

12

study is planning 5-yearly, longitudinal follow up that should clarify: a) whether CSVD

develops more proximally to the expected age of dementia onset in individuals genetically

predisposed to LOAD; b) whether subjects with CSVD in middle age are at increased risk of

subsequently developing LOAD; c) and whether APOE genotype or AD family history interact

with this risk. Ultimately, this will further our understanding of whether CSVD in middle age

might be a prognostic or therapeutic target to reduce dementia risk.

Contributorship statement

JDS and LS processed the data, carried out the statistical analysis, interpreted the data and

drafted the manuscript. EM, NSB and AW processed the data. KW, KR and CWR coordinated

the study and data collection. JOB helped coordinate the study, carried out the statistical

analysis, interpreted the data and drafted the manuscript. All authors revised the

manuscript and gave final approval for the article to be published.

Acknowledgements

Funding: The PREVENT Dementia study is funded by the Alzheimer’s Society. West London

Mental Health NHS Trust sponsors the study and the Clinical Trials Facility at the Trust was

the host institution where research assessments were carried out. MRI scans were

completed at the Imperial College London Clinical Imaging Facility. JS, LS, EM, NSB and JOB

are supported by the NIHR Biomedical Research Unit in Dementia and the Biomedical

Research Centre awarded to Cambridge University Hospitals NHS Foundation Trust and the

University of Cambridge. LS would also like to thank the support of Alzheimer’s Research

UK.

13

Conflicts of interest: none. Sponsors had no role in study design, data collection, data

interpretation, manuscript writing or submission.

References

[1] N.D. Prins, E.J. van Dijk, T. den Heijer, S.E. Vermeer, P.J. Koudstaal, M. Oudkerk, et al.

Cerebral White Matter Lesions and the Risk of Dementia. Arch Neurol 2004;61:1531-1534.

[2] S. Schilling, A.L. DeStefano, P.S. Sachdev, S.H. Choi, K.A. Mather, C.D. DeCarli, et al. APOE

genotype and MRI markers of cerebrovascular disease: systematic review and meta-

analysis. Neurology 2013;81(3):292-300.

[3] S. Lee, F. Vigar, M.E. Zimmerman, A. Narkhede, G. Tosto, T.L. Benzinger, et al. White

matter hyperintensities are a core feature of Alzheimer’s disease: Evidence from the

dominantly inherited Alzheimer network. Ann Neurol 2016;79(6):929-39.

[4] C.W. Ritchie, K. Ritchie. The PREVENT study: a prospective cohort study to identify mid-

life biomarkers of late-onset Alzheimer’s disease. BMJ Open 2012;2:e001893.

[5] M.J. Firbank, T. Minett, J.T. O’Brien. Changes in DWI and MRS associated with white

matter hyperintensities in elderly subjects. Neurology 2003;61:950-954.

[6] M.J. Firbank, L.M. Allan, E.J. Burton, R. Barber, J.T. O’Brien, R.N. Kalaria. Neuroimaging

predictors of death and dementia in a cohort of older stroke survivors. J Neurol Neurosurg

Psychiatry 2012;83:263-267.

[7] S.M. Gregoire, U.J. Chaudhary, M.M. Brown, T.A. Yousry, C. Kallis, H.R. Jager, et al. The

Microbleed Anatomical Rating Scale (MARS): Reliability of a tool to map brain microbleeds.

Neurology 2009;73:1759-1766.

14

[8] M.M.F. Poels, M.W. Vernooij, M.A. Ikram, A. Hofman, G.P. Krestin, A. van der Lugt, et al.

Prevalence and risk factors of cerebral microbleeds: an update of the Rotterdam scan study.

Stroke 2010;41[suppl 1]:S103-S106.

[9] N. Raz, Y. Yang, C.L. Dahle, S. Land. Volume of white matter hyperintensities in healthy

adults: contribution of age, vascular risk factors, and inflammation-related genetic variants.

Biochim Biophys Acta 2012;1822(3):361–369.

[10] R.O. Weller, M. Subash, S.D. Preston, I. Mazanti, R.O. Carare. Perivascular drainage of

amyloid-beta peptides from the brain and its failure in cerebral amyloid angiopathy and

Alzheimer’s disease. Brain Pathol 2008;18(2):253-266.

[11] T. Grimmer, M. Faust, F. Auer, P. Alexopoulos, H. Forstl, G. Henriksen, et al. White

matter hyperintensities predict amyloid increase in Alzheimer's disease. Neurobiol Aging

2012;33(12):2766-2773.

[12] J.M. Olichney, L.A. Hansen, C.R. Hofstetter, J.H. Lee, R. Katzman, L.J. Thal. Association

between severe cerebral amyloid angiopathy and cerebrovascular lesions in Alzheimer

disease is not a spurious one attributable to apolipoprotein E4. Arch Neurol 2000;57(6):869-

874.

[13] M.E. Gurol, A. Viswanathan, C. Gidicsin, T. Hedden, S. Martinez-Ramirez, A. Dumas, et

al. Cerebral amyloid angiopathy burden associated with leukoaraiosis: a positron emission

tomography/magnetic resonance imaging study. Ann Neurol 2013;73(4):529-536.

[14] A. Karch, H. Manthey, C. Ponto, P. Hermann, U. Heinemann, C. Schmidt, et al.

Investigating the association of ApoE genotypes with blood-brain barrier dysfunction

15

measured by cerebrospinal fluid-serum albumin ratio in a cohort of patients with different

types of dementia. PLoS One 2013;8(12):e84405.

[15] M.I. Kester, J.D. Goos, C.E. Teunissen, M.R. Benedictus, F.H. Bouwman, M.P. Wattjes, et

al. Associations between cerebral small-vessel disease and Alzheimer disease pathology as

measured by cerebrospinal fluid biomarkers. JAMA Neurol 2014;71(7):855-862.

[16] P. Alexopoulos, T. Richter-Schmidinger, M. Horn, S. Maus, M. Reichel, C. Sidiripoulos, et

al. Hippocampal volume differences between healthy young apolipoprotein E eta2 and eta4

carriers. J Alzheimers Dis 2011;26(2):207-10.

Footnotes

Abbreviations: CSVD, Cerebral Small Vessel Disease; LOAD, Late-Onset Alzheimer’s Disease;

AD, Alzheimer’s Disease; MRI, Magnetic Resonance Imaging; CMB, Cerebral Microbleed;

SPM, Statistical Parametric Mapping; FLAIR, Fluid-Attenuated Inversion Recovery; MARS,

Microbleed Anatomical Rating Scale; SWI, Susceptibility Weighted Imaging; ICC, Intraclass

Correlation Coefficient; WMH, total White Matter Hyperintensity volume as a percentage of

total brain volume; BBB, Blood-Brain Barrier; CAA, Cerebral Amyloid Angiopathy; CSF,

Cerebrospinal Fluid; PET, Positron Emission Tomography.

16

Figure legend

Figure 1. White matter hyperintensities and cerebral microbleeds

Figure 1a-c (from the same participant): 1a) white matter hyperintensities on FLAIR MRI, 1b)

white matter hyperintensity segmentation produced by SPM8, 1c) final white matter

hyperintensity segmentation after manual image editing. Segmentation enables

quantification of white matter hyperintensity volumes. Figure 1d-e: examples of definite

CMBs on SWI MRI (arrows).

Table legend

Table 1

1a. Comparisons of participant characteristics with WMH

Bivariate analyses of participant characteristics with total white matter hyperintensity

volume as a percentage of total brain volume (WMH) and results of multiple linear

regression of participant characteristics against log10WMH.

1b. Comparisons of participant characteristics with definite CMB presence

Bivariate analyses of participant characteristics with definite CMB presence and results of

multiple logistic regression of participant characteristics against definite CMB presence.

17

18