Migraine and brain lesions. Data from the population-based CAMERA Study Kruit, Mark Christian

Kruit, Mark Christian

Citation

Kruit, M. C. (2010, January 20). Migraine and brain lesions. Data from the population-based CAMERA Study. Department of Radiology, Faculty of Medicine, Leiden University Medical Center (LUMC), Leiden University.

Retrieved from https://hdl.handle.net/1887/14585

Version: Not Applicable (or Unknown) License:

Downloaded from: https://hdl.handle.net/1887/14585

Note: To cite this publication please use the final published version (if applicable).

C H A P T E R  8 

FREQUENTSYNCOPEISARISKFACTORFOR

SUBCLINICALBRAINLESIONS

INBOTHMIGRAINEURSANDCONTROLS

MarkC.Kruit^*

DepartmentofRadiology,LUMC



RolandD.Thijs^*

DepartmentofNeurologyandClinicalNeurophysiology,LUMC



MichelD.Ferrari

DepartmentofNeurology,LUMC



LenoreJ.Launer

LaboratoryofEpidemiology,DemographyandBiometry,

NationalInstituteonAging,NIH,Bethesda



MarkA.vanBuchem

DepartmentofRadiology,LUMC



J.GertvanDijk

DepartmentofNeurologyandClinicalNeurophysiology,LUMC





*Bothauthorscontributedequally.



 Inpreparationforsubmission

A B S T R A C T 

BackgroundandPurposeEarlierstudiesindicatedthatfemalemigraineursareatincreasedriskof

subclinicalbrainlesionsandclinicalstroke.Migraineurshavealsoahigherprevalenceoffrequent

syncope and orthostatic intolerance. The present study assessed whether these items are risk

factorsforbrainlesionsinmigraineursandcontrols.

MethodsMigraineurs (n=291) and controls (n=140) from the populationbased CAMERA cohort

(aged3060years),freeofotherneurologicalsymptoms,underwentabrainMRIscan,astructured

telephoneinterviewincludingquestionsonfrequentsyncope(>5timeslifetime)andsymptomsof

orthostatic intolerance (OI), and a physical examination with blood pressure and heart rate

measurements,assessingorthostatichypotension(OH)andposturaltachycardiasyndrome(POTS).

ResultsAftercontrollingforage,sexandcardiovascularriskfactorsfrequentsyncopeandOIwere

independent risk factors for high load of deep white matter lesions (DWML) in both female

controls (OR=8.8; 95% CI: 1.455) and female migraineurs (OR=5.7; 95% CI:2.115). Female

migraineurs without a history of frequent syncope or OI still ran a significantly increased risk of

highDWMLload (OR=2.4; 95% CI: 1.15.3). Individuals with OI had higher prevalence of high

periventricular WMLload (2% vs. 8%, P=.005), and individuals with OH had higher prevalence of

infratentorial hyperintense lesions load (2% vs. 9%, P=.005), but not of supratentorial lesions.

SyncopeandOIwerenotrelatedtosubclinicalinfarctsorinfratentoriallesions.

ConclusionsFemalecontrolsandmigraineurswithfrequentsyncope,OIorOHareatsignificantly

increasedriskforsubclinicalbrainlesions.

INTRODUCTION

Migraine is a common, multifactorial

neurovascular brain disorder, typically

characterized by recurrent attacks of

disablingheadacheandassociatedsymp

tomsofautonomicdysfunction(migraine

without aura; MO). In up to onethird of

patients, attacks are accompanied by

transient focal neurological dysfunction

(migrainewithaura;MA).^3;203

Whilemigrainecausesacutedistress

and disability during attacks, it is com

monlythoughtnottohaveanylongterm

consequences.However,wefoundinour

populationbasedCAMERAstudy(n=435)

that female migraineurs had a signifi

cantly higher prevalence of deep white

matter hyperintense lesions (DWML),

and that migraineurs had higher preva

lences of infratentorial lesions, including

hyperintense brainstem lesions and

cerebellar infarcts.^71;201 Cardiovascular

riskfactorsandantimigrainemedication

did not modify the association between

the structural brain changes and mi

graine.Fromaclinicalpointofviewthese

MRI lesions appear to have developed

silently,buttheymightstillberelevantin

view of the findings of a a recent meta

analysis, showing that migraine is a risk

factor for ischemic stroke.⁶⁵ The mecha

nisms leading to increased risk for MRI

lesions or stroke in migraineurs remain

unknown.

In the same cohort we examined

clinical presentations of autonomic

nervous system (ANS) dysfunction, as

well as blood pressure (BP) and heart

rate (HR) reflex tests. We found an in

creased lifetime prevalence of syncope,

frequentsyncopeandorthostaticintoler

ance (OI) in migraineurs compared to

controls, without clear interictal signs of

ANS failure.²²⁴ Abnormal reactivity of

blood pressure or heart rate, i.e., or

thostatichypotension (OH)²²⁵orpostural

tachycardia syndrome (POTS),²¹⁸ was not

found more often in migraineurs. It

should be understood that such meas

urements,takenonarandomday,donot

exclude that patients may suffer from OI

and(frequent)syncope.

Although syncope causes per defini

tion cerebral hypoperfusion, it has not

been studied yet whether frequent

syncope is harmful to the brain. In pa

tients with dementia with Lewy bodies

(DLB) the magnitude of fall of systolic

blood pressure during carotis sinus mas

sage (CSM) correlated with the severity

of DWML.²²⁶ Since a hypotensive re

sponse to CSM is a risk factor for syn

cope, these findings may indicate that

syncopecontributestoDWMLinpatients

with DLB. Alternatively, the hypotensive

response to CSM may be a surrogate

markerofdiseaseseverityandtheDWML

are explained by DLB itself. Few studies

evaluated OH as a risk factor for brain

lesions.^227;228Thelongitudinalanalysesof

theARICstudyidentifiedOHaspredictor

ofischemicstroke,evenafteradjustment

for other risk factors.²²⁸ However, these

analyses were not controlled for the

presenceofmigraine.

In this study we assessed whether

frequent syncope is an independent risk

factor for brain lesions in healthy con

trols. In addition, given the increased

prevalenceofbothbrainlesionsandANS

symptoms in migraine, we studied

whether these items could explain (part

of) the increased risk of white matter

lesionsinfemalemigraineurs.

METHODS

STUDYPOPULATION

The Dutch populationbased GEM study

and its substudy (the CAMERA study)

have been described previously.^7;71 In

brief, 863 migraineurs and 5628 controls

wereidentifiedintheGEMstudyaccord

ing to the IHScriteria.³ From those aged

3060 yrs we randomly selected 134

migraineurswithoutaura(MO),161with

aura (MA) and 140 controls without

headaches, frequency matched by sex,

five year age strata and place of resi

dence for the MRI study. Cases and

controls did not differ in age composi

tion, sex proportions, and cardiovascular

risk factors (i.e., body mass index, smok

ing, cholesterol, BP, diabetes and oral

contraceptive use); this also held for or

responders(69%)vs.nonresponders.

Participants gave written informed

consent and participated without any

financial reimbursement. The study

protocol was approved by the ethics

committee and included a telephone

interview and a clinic visit for blood

drawn, a physical and neurological ex

amination and a brain MRI study. The

interviewwascarriedoutbyoneofthree

trained interviewers, who used a com

puterized, structured interview, with

relevant questions presented on the

screenthathadtobereadaloudliterally.

The hospital visit took place within 10

daysofthetelephoneinterview.

CONFOUNDERS,COVARIATESAND

MIGRAINECHARACTERISTICS

Sociodemographic, medical and migraine

characteristics were assessed by inter

view and physical examination, and are

described in detail elsewhere.⁷¹ BP was

noted as the mean of three measure

mentsobtainedintheupperarmwithan

electronic BP monitor. Hypertension was

definedasasystolicBP160mmHgora

diastolic BP 95 mmHg or current use of

antihypertensive drugs. Migraine cases

estimated headache and aura attack

frequency,andfrequencyandamountof

specific antimigraine medication (ergo

tamines, triptans) they used in the years

theyhadmigraineattacks.⁷¹

SYNCOPERELATEDVARIABLES

In the interview, syncope (‘fainting’) was

explainedasabrieflossofconsciousness

that may be provoked by the sight of

blood or while standing for a long time,

butthatcanoccurwithoutaclearprovo

cation. Unconsciousness due to head

injury or an epileptic seizure were ex

plained as entities that should not be

counted as syncope. The occurrence of

syncope was defined twice, first as at

least one attack (‘syncope ever’) and

secondly as ‘frequent syncope’, i.e. five

or more syncopal attacks. The question

naire also investigated the prevalence

and the occurrence of symptoms (syn

cope, near syncope or avoidance) during

specific circumstances, including pro

longed standing, long hot showers, after

heavymeals,afterexerciseoratthesight

ofbloodorduringavenipuncture.These

circumstances are known syncope trig

gers for patients with autonomic failure,

vasovagal syncope or OI, with as excep

tion that venipuncture only acts as a

syncope trigger for patientswith vasova

gal syncope and a heavy meal only for

some patients with autonomic fail

ure.^203;215'Presyncope'wasdefinedasthe

symptoms which usually precede a faint

but can also occur separately, consisting

of dizziness, lightheadedness, loss of

concentration, ringing in the ears or

darkeningofsight.

Twoautonomicreactivitytestswere

performed: change of BP and HR follow

ing standing up and following venipunc

ture. BP and HR measurements were

made with the subject sitting, after 5

minutes supine rest,during 1, 2, 3 and 4

minutes standing, and directly after

venipuncture. The average of the meas

urements while standing was defined as

'standing BP' or 'standing HR'. The maxi

mal decrease while standing was calcu

lated by subtracting the lowest BP value

while standing from the resting BP.

Similarly,themaximalincreaseinHRwas

calculated.OrthostaticchangesofBPand

HR were measured in all but 10 partici

pants (8 migraineurs, 2 controls). Mi

graine cases underwent all examinations

inaheadachefreeperiod(3daysaftera

migraineattack).

COMBINEDENDPOINTS

OH was defined as a fall in SBP 20

mmHg or DBP 10 mmHg within three

minutes after standing up.²²⁵ OI was

definedastheprovocationofsyncopeor

presyncope upon standing or the avoid

ance of prolonged standing.²²⁴ POTS was

defined as OI accompanied by a HR

increase of >30 beats per minute (bpm)

orbyaHR>120bpmwithin5minutesof

standing.²¹⁸

MRI

Whole brain MRI images were acquired

on a 1.0 T and a 1.5T unit with 48 con

tiguous3mmaxialslices.Protocolsinthe

two centers were comparable, including

a combined proton density (PD) and T2

weighted fast spinecho and fluid

attenuated inversionrecovery sequence.

Blinded to clinical status, one neurora

diologist read all images and systemati

cally recorded topographic details of

periventricular and deep white matter

lesion (PVWML and DWML), infraten

torial hyperintense lesions (IHL) and

infarcts. A complete description of the

WMLratingmethodshasbeendescribed

previously,⁷¹ and was based on semi

quantitative scales with known reliability

and validity.¹¹⁸ PVWML scores >3 [out of

15]andthehighest20%volumesoftotal

DWML load were classified in the analy

ses as high load. IHL were hyperintense

on PD and T2 images, and were not

hypointense on FLAIR images.^172;173 Infarctsweredefinedasnonmassparen

chymal defects, with a vascular distribu

tion, isointense to CSFsignal on all se

quences, and, when supratentorial,

surrounded by a hyperintense rim on

FLAIRandPDimages.^113;114

STATISTICS

F² tests, unpaired ttests and oneway

analysesofvariancewereusedtotestfor

any differences in the distributions and

means of measured characteristics

amongthestudygroups.Theassessment

of associations of migraine and its sub

types (MO/MA) to the risk of infarcts,

highPVWML, highDWMLload and IHL

was described inprevious reports.^71;201 It

wasbasedonlogisticregressionanalyses

controlling for age, sex, cardiovascular

riskfactorsandantimigrainemedication.

Analyses were stratified by sex because

of the previously described difference in

prevalence and risk⁷¹ and the reported

difference in prevalence and risk of

stroke.^58;62

The prevalence comparison of syn

copevariables between migraineurs and

controls was described previously.²²⁴ As

thatgroupalsoincludedsubjectswithout

MRIdatawenowrecalculatedthestatis

tics for the current subgroup of subjects

withMRIdata(n=431).

The association between brain le

sions and syncoperelated variables was

assessedbycomparingpresence(yes/no)

of infarcts, highDWMLload, high

PVWMLload and IHL vs. preselected

syncoperelated variables (syncope ever,

frequent [>5x] syncope, OI, and OH).

Using logistic regression analysis, we

examinedtherisk(OddsRatios[OR],95%

confidence intervals [CI]) for the MRI

outcomes by syncopevariables, control

ling for age, sex and municipality, and

stratified these analyses in addition by

sex and migraine status (controls and

migraineursseparately).

Finally, based on the previous find

ings of increased risk of infratentorial

lesions in migraineurs and notably the

increased risk of high DWML load in

femalemigraineursontheonehand,and

on the other hand the current finding of

increased risk of brain lesions in those

withpositivesyncopevariables(frequent

syncope,OIandOH),wetestedforstatis

tical interaction between migraine and

each of these syncope variables with

respect to the risk of brain lesions. This

wasdonebyaddingacrossproductterm

to the model (model 1; controlling for

age, sex, municipality), i.e. migraine x

‘syncope variable’. Because of small

numbers, this was possible only for the

DWML outcome. These analyses were

explicitly also stratified by sex (because

theestablishedassociationbetweenhigh

DWML load and migraineurs was only in

female),andadditionallyORswerecalcu

lated after controlling for low education,

body mass index, hypertension, choles

terol, alcohol use categories, and smok

ing(model2).⁷¹

Analysesandappropriateregression

diagnostics were conducted with the

Statistical Package of Social Sciences

(version10.0.5;SPSSInc.,Chicago,IL).

RESULTS

Syncope data were available from 140

controls and 291 migraineurs out of 435

participants with complete MRI data.

Syncope data were lacking in 4 partici

pants,whodidnotappeartodifferfrom

the remainder with respect to demo

graphic variables, cardiovascular risk

factors,orMRIvariables.TABLE1showsa

higher prevalence of syncope, frequent

syncope and OI in migraineurs vs. con

trols, confirming our earlier findings in

TABLE 1 Prevalence of syncope, orthostatic intolerance and orthostatic hypotension in

participantsofthepopulationbasedCAMERAMRIstudy(n=431)

 Controls  Migraineurs  Pvalue

SYNCOPEEVER  43/140 31%   144/291 49% <.0001

 Male  10/40 25%   27/75 36% .2

 Female  33/100 33%   117/216 54% <.0001

FREQUENT(>5X)SYNCOPE  8/137 6%   40/282 14% .01

 Male  2/40 5%   1/73 1% .2

 Female   6/97 6%   39/209 19% .004

ORTHOSTATIC

INTOLERANCE

 17/140 12%   94/290 32% <.0001

 Male  2/40 5%   10/75 13% .16

 Female  15/100 15%   84/215 39% <.0001

ORTHOSTATIC

HYPOTENSION

 20/138 14%   47/287 16% .6

 Male  6/40 15%   11/75 15% .9

 Female  14/98 14%   36/212 17% .5

POTS  3/138 2%   9/285 3% .6

 Male  0/40 0%   3/74 4% .2

 Female  3/98 3%   6/212 3% .9

Valuesareexpressedasnumbersandproportions(%).

PvaluesarefromPearson's2test(unadjusted).

this smaller subset of participants.²²⁴ Afterstratificationbysex,thedifferences

remained significant only in female

participants, but it should be noted that

thenumberofmenwasmuchlowerthan

that of women. No differences were

found between migraineurs with and

without aura, nor between those with

fewer (<1 attack/month) and more at

tacks. There was no difference in preva

lence of OH or POTS. Diabetes mellitus,

history of myocardial infarction, BMI,

highalcoholintake,anduseofantihyper

tensives, were found negative as poten

tialriskfactorsforsyncope.

TABLE 2 shows the prevalence of high

DWMLload,highPVWMLloadandIHLin

those with and without history of fre

quent syncope, OI and OH (crude data).

No differences in prevalence of lesions

were found for those with vs. those

withouthistoryofsyncopeever,orPOTS.

Thosewithahistoryoffrequentsyncope

had higher prevalence of highDWML

TABLE 2 Prevalence of syncope, orthostatic intolerance and orthostatic hypotension vs.

presenceofMRIhyperintenselesions(n=431)

 Frequentsyncope  Orthostaticintolerance  Orthostatichypotension

 no yes Pvalue  no yes Pvalue  no yes Pvalue

ALL n=371 n=48   n=319 n=111   n=358 n=67

 HighDWMLload 18% 35% .005  17% 28% .01  20% 19% .9

 HighPVWMLload 3% 8% .08  2% 8% .005  4% 3% .7

 IHL 3% 2% .7  2% 5% .1  2% 9% .005

FEMALE n=261 n=45  n=217 n=99  n=260 n=50 

 HighDWMLload 17% 38% .002  17% 27% .03  20% 18% .7

 HighPVWMLload 3% 9% .06  2% 7% .04  4% 2% .4

 IHL 3% 2% .8  2% 5% .2  2% 10% .003

MALE n=110 n=3  n=103 n=12  n=98 n=17 

 HighDWMLload 20% 0% .4  18% 33% .2  19% 23% .7

 HighPVWMLload 4% 0% .7  2% 17% .008  4% 0% .7

 IHL 4% 0% .7  3% 8% .4  3% 6% .6

MIGRAINEURS n=242 n=40  n=197 n=95  n=240 n=47 

 HighDWMLload 20% 35% .04  20% 27% .2  22% 23% .8

 HighPVWMLload 4% 5% .7  3% 6% .1  4% 4% .9

 IHL 4% 2% .6  4% 5% .6  3% 13% .003

         

CONTROLS n=129 n=8   n=123 n=1   n=118 n=20 

 HighDWMLload 14% 37% .07  13% 35% .02  17% 10% .4

 HighPVWMLload 2% 25% .001  2% 18% .001  4% 0% .4

 IHL 1% 0% .8  0% 6% .007  1% 0% .7

PvaluesarefromPearson's²test(unadjusted).HighDWMLloadreflectstheupper20^thpercentileofthetotal

DWML volume distribution. HighPVWMLload is based on a score of >3 [out of 15] for PVWML. IHL is the

presenceofinfratentorial(brainstemand/orcerebellum)hyperintenselesion(s).

load. After stratification for sex this was

the case only for women, not men; and

after stratification by migraine diagnosis,

this held for both migraineurs and con

trols (trend in controls, P=.07). After

controlling for age, municipality and sex

(Model1),theseeffectsremainedforthe

risk of highDWMLload by frequent

syncope: OR=2.8 (95% CI 1.45.6) for all

subjects, OR=3.3 (95% CI 1.66.8) for

female, OR=2.3 (95% CI 1.05.0) for

female migraineurs, and OR=23 (95% CI

1.8297) for female controls. A trend

towards a higher prevalence of high

PVWMLload was present in females,

which seems to be limited to controls.

We found no difference with respect to

the prevalence of IHL for those with vs.

withoutfrequentsyncope.

ThosewithsymptomsofOIalsohad

a higher prevalence of highDWMLload,

again most explicitly in females; the

effects remained also after controlling

(Model 1): OR=2.0 (95%CI 1.23.5) for all

subjects; OR=2.0 (95%CI 1.13.7) for

female;OR=7.4(95%CI1.248)forfemale

controls. In addition, in those with OI

there was also a higher prevalence of

highPVWMLload, in both female and

male, and with largest effect in controls

(n=1).OHwasnotrelatedtohighDWML

load or highPVWMLload. However,

TABLE 3RiskofhighDWMLloadinrelationtomigraineandcoexistentsyncoperelated

riskfactors(n=431)

 All Female

  Model1  Model2  Model1  Model2



no.* OR(CI)

P

value  OR(CI)

P

value no.* OR(CI)

P

value  OR(CI)

P

value

FREQUENT (>5X) SYNCOPE

        

 controlswithout† 18/129 1  1  9/91 1  1 

 migraineurswithout 49/242 1.5(0.82.7) .2 1.4(0.22.7) .2 36/170 2.4(1.15.3) .04 2.3(1.05.3) .05

 controlswith 3/8 3.9(0.819) .1 3.5(0.718) .1 3/6 8.8(1.455) .02 6.3(1.040) .05

 migraineurswith 14/40 3.6(1.58.7) .004 3.5(1.48.7) .006 14/39 5.7(2.115) .001 6.0(2.117) .001 ORTHOSTATIC

INTOLERANCE

        

 controlswithout† 16/123 1  1  9/85 1  1 

 migraineurswithout 39/197 1.6(0.83.1) .2 1.4(0.72.8) .3 27/132 2.2(1.05.1) .06 2.0(0.84.7) .1

 controlswith 6/17 4.3(1.314) .02 4.3(1.215) .02 4/15 3.9(1.016) .06 3.9(0.917) .07

 migraineurswith 25/95 2.4(1.25.1) .02 2.4(1.15.2) .03 23/84 3.4(1.48.1) .006 3.5(1.48.6) .008

Values are numbers, odds ratios, 95% confidence intervals and Pvalues. HighDWMLload reflects the upper

20^th percentile of the total deep white matter lesionvolume distribution. Odds ratios (95% CI, confidence

interval) have been calculated by logistic regression, controlling for age, sex, municipality (model 1), and low

education,bodymassindex,hypertension,cholesterol,alcoholusecategories,andsmoking(model2).

* NumberofsubjectswithhighDWMLload/totalnumberofsubjectsinthissyncopesubgroup.

† Referencegroup.

those with OH had significantly more

often IHL, most explicitly in female (and)

migraineurs.

We found no difference in preva

lence of brain infarcts in those without

vs. those with frequent syncope (8% vs.

2%;P=.1),OI(8%vs.5%;P=.4),orOH(7%

vs. 7%; P=.9); this did not change after

stratification by sex, migraine status or

migrainesubtype.Similarly,therewasno

differenceintheprevalenceofcerebellar

infarcts.

Becausewefoundbothasignificant

association between migraine and syn

cope variables, and also between syn

cope variables and brain lesions, we

tested for statistical interaction. Interac

tion was proven for 'migraine x frequent

syncope' (P=.01) and 'migraine x OI'

(P=.08) vs. highDWMLload. Using

dummyvariables,wethencalculatedthe

independent risk factors for each combi

nation of 'diagnosis x syncope status'

(TABLE3),stratifiedforsex.Thisrevealed

an independently increased risk of high

DWMLload in female migraineurs with

out history of frequent syncope (OR 2.4,

P=.04),infemalecontrolswithhistoryof

frequent syncope (OR 8.8, P=.02), and

also in female migraineurs with frequent

syncope (OR 5.7, P=.001). Similarly, for

female subjects with and without mi

graine, OI was found to be a risk factor

for highDWMLload. Additional calcula

tions (Model 2) after entering cardiovas

cular risk factors in the model, showed

similar results, confirming the independ

encyoftheassessedvariables.

DISCUSSION

We present evidence that syncope

related ANS symptoms are a risk factor

for highDWMLload and highPVWML

load, with strongest effects in women,

both with and without migraine. Among

female migraineurs,both those with and

without frequent syncope or OI were at

increased risk of highDWMLload.

Among female controls those with fre

quentsyncopeorOIarealsoatincreased

riskforhighDWMLload,whereasOIalso

seems to predispose to highPVWML

load, and OH to IHL. The findings for

migraineurs and controls will be sepa

ratelydiscussedbelow.

SYNCOPEASARISKFACTORFOR

BRAINLESIONSINCONTROLS

Syncope is basically defined as transient,

brief and selflimited loss of conscious

nessduetocerebralhypoperfusion,with

spontaneous recovery.²⁰³ There are

variousgroupsofcauses,includingreflex

syncope (synonym: neurally mediated

syncope), cardiac syncope and or

thostatic hypotension. Whereas long

lasting cerebral hypoperfusion can cer

tainly cause brain damage, shortlasting

hypoperfusioninsyncopeiscurrentlynot

thought to cause any permanent dam

age, not even when attacks are very

frequent.

WMLs are frequently observed on

brainMRIscansofsubjectsaboveage50,

and their prevalence increases with age.

Their pathogenesis is still incompletely

understood,althoughepisodicorchronic

hypoperfusion, various cardiovascular

risk factors, including BP levels, insuffi

cient bloodbrainbarrier function, al

tered vascular permeability, increased

endothelial activation and microglial

activation due to hypoxic and inflamma

toryinsults,andCSFaccumulationplaya

role.²²⁹ The regions of the white matter

thataresuppliedbythedistalendsofthe

penetratingarterioles(theperiventricular

zones and the semioval center), have in

normal conditions already the lowest

perfusion pressure.²³⁰ These zones are

most vulnerable to periods of relative

hypoperfusion, and perfusion may theo

retically become insufficient (ischemia)

either in conditions of chronically im

paired local cerebral blood flow, as may

happen in subjects with large and/or

small vessel atherosclerosis, or in condi

tions with a paroxysmal reduction in

cardiacoutputorbloodpressure,suchas

insubjectswithOH.

It is therefore conceivable that re

peatedsyncopemaycauseWML.Several

MRI studies indeed found associations

between OH and other dynamic changes

in BP and the risk of WML and/or silent

infarcts. It must be noted that most

studieswereperformedinoldersubjects

with concurrent risk factors like hyper

tensionanddiseaseslikedementia,^226;231

234 meaning that in such subjects both

local factors and systemic hypotension

may have colluded to result in WMLs.

StrongepidemiologicevidencethatOHis

an independent predictor of ischemic

stroke in middleaged persons comes

fromthelongitudinaldatafromtheARIC

study.²²⁸

We now showed with MRI that a history

of frequent syncope or of OI is an inde

pendent risk factor for highDWMLload,

and likely also for highPVWMLload and

IHL, in healthy young to middleaged

subjects from the general population.

Effects were strongest in women, and

similar for females with and without

migraine (both OR=6; Model 2). These

results are in line with the above de

scribed findings from other studies, and

are possibly based on the same patho

physiological mechanisms as in OH, that

likely all lead to temporary cerebral

hypoperfusion.²²⁸ It should be stressed

howeverthatwestudiedsubclinicalbrain

lesions; the clinical relevance of these

lesions in healthy controls with frequent

syncopeisunknown.

Multivariate analyses allowed us to

control the associations for the possible

confoundingeffectsofcardiovascularrisk

factors, including diabetes, hypertension

and antihypertensive treatment. Al

though there were no participants with

overt heart disease, we cannot exclude

thepossibilitythatsomesyncoperelated

variablesmightactasaproxyforcardio

vascular risk factors.²²⁸ The relatively

young age of the study group suggests

that this explanation is not likely. Since

data were acquired blinded for migraine

diagnosis, the stratification by diagnosis

(into migraineurs and controls sepa

rately) is also not a likely source of bias.

Even if migraineurs would for unknown

reasonshavebeenmorelikelytoanswer

syncope questions affirmatively, this

would not explain any relationship be

tween syncope and DWML in controls.

The finding that the association of fre

quent syncope andOI to DWML reached

a higher level of significance in women

than in men probably represents a sam

ple size effect, as there were more

women,incommonwiththesexdistribu

tionofmigraine(female:male~3:1).

SYNCOPEASARISKFACTORFOR

BRAINLESIONSINMIGRAINEURS

The second purpose of this study was to

address whether ANS dysfunction in

migraineurscouldexplaintheirincreased

risk of brain lesions. We unraveled this

complex relationship using multivariate

analyses, and found that female mi

graineurs with frequent syncope or OI

wereatathreetimeshigherriskofhigh

DWMLload compared with female

migraineurs without such history. How

ever, female migraineurs without syn

cope still were at significantly increased

riskcomparedtofemalecontrolswithout

syncope. We therefore conclude that

frequent syncope or OI does not explain

the increased risk of DWML lesions in

female migraineurs on its own. Instead,

both frequent syncope and OI as well as

migraineincreasetheriskofsuchlesions.

The precise pathophysiological mecha

nismsunderlyingthesecomplaintsareas

yet unknown and require further study.

Although our data indicate an contrib

utive role of ANSsymptoms in the etiol

ogy of WMLs, e.g. longitudinal data first

have to prove that lesion load increases

withongoingANSsymptomsandsigns.

There is evidence that the sympathetic

nervoussystemmayexhibitsubtlealtera

tions in migraineurs.²²³ These include

reduced supine plasma norepinephrine

levelsandperipheraladrenergicreceptor

supersensitivity; the author proposed

subtle functional disturbances in an

anatomicalllyintactsystem.²²³

STUDYSTRENGTHSAND

LIMITATIONS

Strengths of this study have been de

tailed and confirmed earlier.^71;235 For the

currentanalyses,thedetaileddescription

of the cohort allowed us to control for

possibly confounding factors due to

cardiovascular disease or medication

effects.WeusedstandardizedBPandHR

measurements, that were obtained

blindedfor(migraine/syncope)diagnosis.

Although our measures of syncope,

frequent syncope and OI were historical,

andthuspotentiallysubjecttorecallbias,

we are confident that by using relative

robust cutoff values for e.g. ‘frequent

syncope’, and the use of different ques

tions to assess orthostatic complaints,

thesedataproducedavalidclassification

ofindividualswithorwithoutANSsymp

toms. Because of logistical reasons, we

wereunabletoperformamoreextensive

ANS test battery, e.g. with tilttable

testing; with the current evidence such

assessments will likely allow for a more

precise evaluation of subtypes of ANS

dysfunction.

In summary, we identified frequent

syncope and OI as independent risk

factors for highDWMLload in both

controls andmigraineurs,with the effect

beingstrongestinwomen.Thesefindings

suggest that those with impaired control

of cerebral perfusion pressure, are at

increased risk for ischemic white matter

tissue damage. In female migraineurs,

this adds to the known increased risk of

highDWMLload, but does not com

pletely explain it. Impaired SNS function

mightbe considered as a possible key to

understand the described associations.

However, additional clinical and epide

miological studiesare needed toconfirm

and explore these findings. Assessment

of parameters of orthostatic propensity

can be done by clinical history, and

probably adds relevant information to

stroke and/or cerebrovascular risk profil

ing both in clinical and epidemiological

settings.