Recurrentstrokeriskandcerebralmicrobleed burden in ...burden in ischemic stroke and TIA A...

Duncan Wilson MDAndreas Charidimou

PhDGareth Ambler PhDZoe V Fox PhDSimone Gregoire PhDPhillip RaysonToshio Imaizumi MDFelix Fluri MDHiromitsu Naka MDSolveig Horstmann MDRoland Veltkamp MDPeter M Rothwell PhDVincent IH Kwa MDVincent Thijs MDYong-Seok Lee MDYoung Dae Kim MDYining Huang MDKa Sing Wong MDHans Rolf Jaumlger MD

FRCRDavid J Werring PhD

Correspondence toDr Werringdwerringuclacuk

Supplemental dataat Neurologyorg

Recurrent stroke risk and cerebralmicrobleedburden in ischemic stroke and TIAA meta-analysis

ABSTRACT

Objective To determine associations between cerebral microbleed (CMB) burden with recurrentischemic stroke (IS) and intracerebral hemorrhage (ICH) risk after IS or TIA

Methods We identified prospective studies of patients with IS or TIA that investigated CMBs andstroke (ICH and IS) risk during $3 months follow-up Authors provided aggregate summary-leveldata on stroke outcomes with CMBs categorized according to burden (single 2ndash4 and$5CMBs)and distribution We calculated absolute event rates and pooled risk ratios (RR) using random-effects meta-analysis

ResultsWe included 5068 patients from 15 studies There were 1151284 (96) recurrent ISevents in patients with CMBs vs 2123781 (56) in patients without CMBs (pooled RR 18 forCMBs vs no CMBs 95 confidence interval [CI] 14ndash25) There were 491142 (43) ICHevents in those with CMBs vs 172912 (058) in those without CMBs (pooled RR 63 for CMBsvs no CMBs 95CI 35ndash114) Increasing CMB burden increased the risk of IS (pooled RR [95CI] 18 [10ndash31] 24 [13ndash44] and 27 [15ndash49] for 1 CMB 2ndash4 CMBs and $5 CMBs respec-tively) and ICH (pooled RR [95 CI] 46 [19ndash107] 56 [24ndash133] and 141 [69ndash290] for 1CMB 2ndash4 CMBs and $5 CMBs respectively)

Conclusions CMBs are associated with increased stroke risk after IS or TIA With increasing CMBburden (compared to no CMBs) the risk of ICH increasesmore steeply than that of IS However ISabsolute event rates remain higher than ICH absolute event rates in all CMB burden categoriesNeurologyreg 2016871501ndash1510

GLOSSARYCAA 5 cerebral amyloid angiopathy CI5 confidence interval CMB 5 cerebral microbleed ICH 5 intracerebral hemorrhageIS 5 ischemic stroke SWI 5 susceptibility-weighted imaging

Cerebral microbleeds (CMBs) are radiologically defined small round or ovoid regions of signalloss seen on paramagnetic MRI sequences1 In the limited available pathologic correlationstudies CMBs mostly correspond to hemosiderin-laden macrophages close to vessels affectedby small vessel disease2ndash5 It is thus inferred that CMBs are a marker of direct extravasation oferythrocytes from arterioles and capillaries damaged by bleeding-prone arteriopathies An arte-riopathy associated with systemic arterial hypertension and pathologic changes in small perfo-rating arteries of the deep gray and white matter causes CMBs in deep (basal ganglia) as well aslobar regions In Western (Caucasian) people with intracerebral hemorrhage (ICH) CMBs in

From the Stroke Research Centre (DW AC SG PR DJW) Department of Brain Repair and Rehabilitation UCL Institute ofNeurology and The National Hospital for Neurology and Neurosurgery Department of Statistical Science (GA) and Biomedical ResearchCentre (ZVF) UCL London UK Department of Neurosurgery (TI) Kushiro City General Hospital Hokkaido Japan Department ofNeurology (FF) University Hospital Wuumlrzburg Germany Department of Neurology (HN) Suiseikai Kajikawa Hospital Hiroshima JapanDepartment of Neurology (SH) University of Heidelberg Germany Department of Stroke Medicine (RV) Division of Brain SciencesImperial College London Nuffield Department of Clinical Neurosciences (PMR) John Radcliffe Hospital University of Oxford UKDepartment of Neurology (VIHK) Onze Lieve Vrouwe Gasthuis Amsterdam the Netherlands Department of Neurology (VT) AustinHealth and Melbourne Brain Center University of Melbourne Australia Department of Neurology (Y-SL) Seoul National UniversityBoramae Medical Center Department of Neurology (YDK) Yonsei University College of Medicine Seoul Korea Department of Neurology(YH) Peking University First Hospital Beijing China Division of Neurology (KSW) Department of Medicine amp Therapeutics TheChinese University of Hong Kong Prince of Wales Hospital Shatin Hong Kong and Lysholm Department of Neuroradiology (HRJ)National Hospital for Neurology and Neurosurgery London UK

Go to Neurologyorg for full disclosures Funding information and disclosures deemed relevant by the authors if any are provided at the end of the articleThe Article Processing Charge was paid by Wellcome Trust

This is an open access article distributed under the terms of the Creative Commons Attribution License 40 (CC BY) which permits unrestricteduse distribution and reproduction in any medium provided the original work is properly cited

copy 2016 American Academy of Neurology 1501

ordf 2016 American Academy of Neurology Unauthorized reproduction of this article is prohibited

a strictly lobar distribution are highly specificfor cerebral amyloid angiopathy (CAA) whichcauses progressive deposition of b-amyloid insmall cortical and leptomeningeal arterialwalls6 though this pattern may not be so spe-cific in Eastern (Asian) people7 and in thosewithout ICH8

Multiple prospective studies in ischemicstroke (IS) cohorts have shown that CMBsare associated with subsequent ICH risk910

However CMBs are also associated withincreased subsequent IS risk11ndash14 Indeed sug-gested ischemic mechanisms for CMBsinclude ischemia-mediated iron store releaseby oligodendrocytes15 phagocytosis of red cellmicroemboli into the perivascular space(termed angiophagy)16 or hemorrhagic trans-formation of small microinfarcts17 Indeed ina recent community study after adjusting forcardiovascular risk factors CMBs were foundto be associated with lacunes and white mattervolume progression18 Few data are availableon how CMB burden affects the balance ofICH and IS risk in different populations InCAA cohorts an increasing number of CMBsis associated with an increased risk of ICHsuggesting a relationship between CMB num-ber and the severity of bleeding-prone arterio-pathy19 Whether an increasing number ofCMBs is also associated with an increased riskof ICH in IS and TIA cohorts remains uncer-tain If increasing CMB burden shifts the bal-ance of risk toward ICH rather than IS thiscould have major clinical relevance for antith-rombotic risk-benefit decisions after IS andTIA Our previous meta-analysis of 10 pro-spective studies including 3067 patients withIS or TIA found that CMB presence is associ-ated with a higher risk of ICH than IS20 (theodds ratio was 853 for ICH and 155 for IS)but was not able to address the key clinicalquestion of how the number (burden) ofCMBs influences ICH and IS risk

We therefore performed a pooled analysis ofaggregate summary data including CMB bur-den and distribution to investigate the risk ofsubsequent IS and ICH in individuals who havehad an IS or TIA We tested the followinghypotheses (1) CMB presence is associatedwith an increased risk of stroke (ICH IS)and (2) as CMB burden increases (due to amore

severe bleeding-prone arteriopathy) the risk ofICH increases more steeply than the risk of IS

METHODS We searched Medline and Embase from 1996 (the

year CMBs were first reported) through to April 2015 Our

search strategy was as follows

1 ldquoCerebral microbleedrdquo or CMB or ldquocerebral microhemorrrdquo

or ldquobrain microbleedrdquo or ldquobrain microhemorrrdquo

2 Stroke or ldquoischemic strokerdquo or TIA or ldquointrac adj2 h

emorrhagrdquo or ICH

3 1 and 2

We included published and unpublished studies fulfilling the

following criteria (1) performed paramagnetic-sensitive MRI se-

quences to detect CMBs at baseline (2) assessed CMBs at base-

line and associations with IS or ICH as primary or secondary

outcomes (3) had a prospective study design with at least 3

months of follow-up and (4) fulfilled at least 4 of 6 predefined

quality indicators We excluded cross-sectional studies and case

series Two clinical research fellows (AC and DW) reviewed

each study for eligibility

Data extraction We contacted all authors to provide data

on study population size patient-year follow-up and

antithrombotic treatment We obtained data on outcome

events of symptomatic IS and ICH with baseline CMB number

categories as follows CMB present 1 CMB 1 CMB 2ndash4

CMBs 5ndash10 CMBs 10 CMBs strictly deep CMBs strictly

lobar CMBs and mixed distribution CMBs We extracted

all demographic imaging and follow-up outcome data

from each study

Quality assessment and reducing the risk of bias All

included studies were critically appraised against a checklist of 6

key quality indicators (table e-1 at Neurologyorg) with

reference to the Strengthening the Reporting of Observational

Studies in Epidemiology statement and the Preferred Reporting

Items for Systematic Reviews and Meta-Analyses guidelines The

quality criteria included assessment for bias and all studies had

a quality score of $46

Statistical methods We first performed separate random effect

meta-analyses to derive summary estimates of the pooled risk

ratios of ICH and IS for each CMB category vs the reference

category of ldquono CMBsrdquo Due to the small number of events in

some studies the categories 5ndash10 and 10 were combined in

line with previous studies of CMB burden and prognosis which

demonstrated their prognostic relevance for future ICH risk19

Logistic regression was then used to estimate the increased risk

(odds) of ISICH for each additional CMB First CMB

categories were converted to a continuous scale by assuming

that patients in the CMB groups 0 1 2ndash4 5ndash10 and 10

had on average 0 1 3 75 and 125 CMBs respectively

Then for each study a logistic regression model was fitted

relating the (log) odds of IS or ICH to the (estimated) number

of CMBs These (log) odds ratios were then pooled using random

effects meta-analysis

We calculated the I2 statistic to investigate heterogeneity

Funnel plots (Begg and Mazumdar) were generated to investigate

publication bias Finally where necessary we undertook meta-

regression of confounding covariates of biological plausibility or

with differences between the studies (average follow-up age

hypertension prevalence demographics antithrombotic use)

All analyses were performed using STATA 120 (StataCorp LP

College Station TX)

1502 Neurology 87 October 4 2016


Standard protocol approvals registrations and patientconsents Individual studies and data transfer protocols were

approved by local ethics committees No additional ethical

approval was required for this meta-analysis

RESULTS Fifteen studies met our inclusion criteria(12 published and 3 unpublished) including 5068patients911ndash1421ndash27 (figure 1) The patient and datacharacteristics of each study are shown in tables e-2and e-3 Twelve of the 15 studies provided data onCMBs at the time of the initial ISTIA fully stratifiedinto number categories and location Eight studiesinvolving 3111 patients were from predominantlyEastern (Asian) cohorts the remainder (1957 patients)were predominantlyWestern (Caucasian) Two studiesincluded strictly TIA patients 4 included those withTIA or IS and 9 included strictly IS patients Thenumber of patients with CMBs was 1284 giving anoverall pooled prevalence of 253 Median follow-upwas 18 months (interquartile range 11ndash30) Overall79 of patients were prescribed antiplatelet agentsonly 15 of patients were prescribed anticoagulants

(mainly from one study with a high proportionof patients on anticoagulation [87])24 CMBpresence was more prevalent with higher burdenin the Eastern cohorts compared to Western cohorts(table e-3)

CMBs and IS risk The total recurrent IS rate was 3275068 (65) The IS event rate in those with CMBswas 9 (1151284) vs 56 (2123781) for thosewithout CMBs thus CMBs are associated with anabsolute risk increase of 34 for IS The absoluterisk increase for IS for CMBs vs no CMBs increasesas the CMB burden increases (18 for 1 CMB48 for 2ndash4 CMBs and 51 for $5 CMBs[table 1])

The risk ratios for different CMB burden and dis-tribution categories on IS are also shown in table 1The presence of CMBs (vs no CMBs) was associatedwith a pooled risk ratio of recurrent IS of 18 (95confidence interval [CI] 14ndash25) (figure 2) Funnelplots revealed no evidence of publication bias (Eggertest p 5 04) The presence of a single CMB (vs no

Figure 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart of study selection

Neurology 87 October 4 2016 1503


CMBs) had a pooled risk ratio for IS of 18 (95 CI10ndash31) The pooled risk estimates for IS suggest anincreasing trend toward higher IS risk with increas-ing CMB burden (table 1 and figure 3) Logisticregression was used to estimate the increase in riskfor each additional CMB this showed an odds ratioof 110 (95 CI 106ndash114) per CMB increase (fig-ure e-1) The risk estimates for IS for each distribu-tion category of CMBs (vs no CMBs) ranged from16 (95 CI 10ndash27) for strictly deep CMBs to 26(95 CI 15ndash43) for mixed CMBs with overlap-ping confidence intervals for all groups Because wenoted statistical heterogeneity among the cohorts forIS risk (I2 33 47 68 and 51 respectivelyfor CMB presence 1 CMB 2ndash4 CMB and $5CMB respectively) each potential confounder (eth-nicity average follow-up age hypertension preva-lence and antithrombotic use) was investigatedseparately using meta-regression (table e-4 figuree-2) We found only weak evidence for a confound-ing effect of hypertension as expected the relativerisk associated with baseline CMBs on future IS risk(effect size) is attenuated in studies with a higherprevalence of hypertension

CMBs and ICH risk The total ICH rate was 665068(13) The ICH event rate in those with CMBs was43 (491142) vs 05 (172912) for those with-out CMBs thus CMBs confer an absolute riskincrease of 38 for ICH The absolute risk increasefor ICH for CMBs vs no CMBs increases as the CMBburden increases (17 for 1 CMB 18 for 2ndash4CMBs and 82 for $5 CMBs [table 1])

The risk ratios for different CMB burden and dis-tribution categories on ICH are also shown in table 1The presence of CMBs (vs no CMBs) was associatedwith a pooled risk ratio of 63 for subsequent ICH(95 CI 35ndash114) (figure 2) Four studies wereexcluded as they did not report any ICH outcomes

Increasing CMB burden was associated with anincreased risk of ICH (pooled risk ratio 46 [95CI 19ndash107] 56 [95 CI 24ndash133] and 141[95 CI 69ndash290] for 1 CMB 2ndash4 CMBs and$5 CMBs compared to no CMBs respectively)(table 1 and figure 4) Logistic regression showed anodds ratio of 129 (95 CI 121ndash137) for ICH peradditional CMB (figure e-1) Of the CMB anatomi-cal distribution categories (strictly lobar mixed orstrictly deep) strictly lobar CMBs were associatedwith the highest risk of subsequent ICH vs no CMBs(pooled risk ratio 105 [95 CI 45ndash243] table 1)There was no publication bias within studies (Eggertest p 5 098) Meta-regression was not undertakenbecause heterogeneity was not detected (I2 was 0)for ICH outcomes

DISCUSSION Our meta-analysis of 15 prospectivestudies including more than 5000 patientspresenting with IS or TIA found that the presenceof any CMBs is associated with an approximatedoubling of the risk of IS but with anapproximately 6-fold increase in the risk of ICH inkeeping with 2 previous smaller meta-analyses2028

Our meta-analysis also builds on these previousstudies and adds new knowledge on ICH risk firstwe were able to increase our statistical power byincluding more ICH outcomes20 second bypooling aggregate data we investigated howincreasing CMB burden affects the balance betweenfuture IS and ICH (including both relative andabsolute risks) and third we partially adjusted forconfounding factors through meta-regression Ourmost important new finding is that with increasingCMB burden the risk of ICH increases more steeplythan that of IS In patients with$5 CMBs the risk ofICH was substantially higher than that of IS (riskratio for ICH 141 [95 CI 67ndash290] vs risk ratiofor IS 273 [95 CI 15ndash49]) In a complementary

Table 1 Pooled relative risk for recurrent ischemic stroke and intracerebral hemorrhage for different cerebral microbleed (CMB) burden anddistribution (all risk ratios are compared to the reference category of no CMBs)

CMBdistribution n

Ischemic stroke Intracerebral hemorrhage

Pooled absoluteevent rates nN ()

Pooled absoluterisk increase

PooledRR

Lower95 CI

Upper95 CI



PooledRR

Lower95 CI

Upper95 CI

CMB presence 1151284 (9) 34 18 14 25 491142 (43) 38 63 35 114

1 CMB 31433 (72) 18 18 10 31 8354 (23) 17 46 19 107

2ndash4 CMBs 44433 (102) 48 24 13 44 9383 (23) 18 56 24 133

Dagger5 CMBs 34342 (105) 51 27 15 49 24274 (88) 82 141 69 290

Strictly lobar 31332 (93) 39 20 14 29 12332 (36) 32 105 45 243

Strictly deep 29437 (66) 12 16 10 27 6437 (14) 1 33 13 85

Mixed 44411 (107) 53 26 15 43 25411 (61) 57 111 55 226

Abbreviations CI 5 confidence interval n 5 number of events in each subgroup N 5 total number of patients in each subgroup RR 5 risk ratio



logistic regression analysis we showed that eachadditional CMB is associated with an increasedodds of 13 (95 CI 12ndash14) for ICH and 11

(95 CI 11ndash11) for IS supporting a steeperincrease in ICH than IS risk with higher CMBburden However the absolute event rate of

Figure 2 Forest plot of the risk of ischemic stroke (IS) and intracerebral hemorrhage (ICH) for presence of anycerebral microbleeds (CMBs) vs no CMBs

CI 5 confidence interval RR 5 risk ratio



Figure 3 Forest plot of the risk of ischemic stroke (IS) for different burdens of cerebral microbleeds (CMBs) vsno CMBs




Figure 4 Forest plot of the risk of intracerebral hemorrhage (ICH) for different burdens of cerebralmicrobleeds (CMBs) vs no CMBs




recurrent IS was consistently higher than the absoluteevent rate of ICH in patients with CMBs and withinall CMB categories including those with $5 CMBs

A large number of CMBs (eg $5 CMB) mighthelp identify patients at substantially higher risk ofICH than of IS Indeed for clinicians the key ques-tion is what burden of CMBs could tip the balance ofrisk towards ICH sufficiently to affect clinical deci-sions for example antithrombotic drug use Antipla-telet agents only modestly reduce the absolute risk ofIS in secondary prevention (05ndash25)29 Our datashow the absolute risk of ICH increases substantiallymore than the absolute risk of IS as CMB burdenincreases the effect is most evident with $5 CMBswhich is associated with an 82 absolute riskincrease for ICH vs a 51 absolute risk increasefor IS This raises the possibility that antiplatelet drugrisk-benefit assessment may favor avoiding their usein those with numerous CMBs (eg $5 CMBs)This could affect a substantial proportion of IS andTIA patients of patients with CMBs the prevalenceof those with$5 CMBs ranged from 12 to 51 instudies included in this analysis (table e-3) and variedby ethnicity (mean 17 for Western cohorts and35 for Eastern cohorts)

A previous meta-analysis suggested that ethnicitymay be an important determinant of the balance ofICH and IS risks associated with CMBs an increasedrisk for ICH risk was only statistically significant inEastern cohorts while IS risk was only significant inWestern cohorts20 In the present study we includedmore patients and using meta-regression found thatethnicity does not confound the association betweenCMB burden and IS or ICH risks Thus based onthe current study CMB burden appears to bea greater predictor of IS and ICH risk than ethnicity

Our study has a number of strengths includinga large sample size from multiple cohorts from differ-ent countries We only included those of high qualityusing systematic quality indicator assessment Weincluded data on CMB burden and distributionand adjusted for confounding factors through meta-regression Our study thus provides the best currentlyavailable evidence on how CMBs affect IS and ICHrisk after IS or TIA

Our study also has limitations Because weincluded aggregate summary-level data (rather thanindividual patient data) we could not explore theeffect of CMB distribution free from the confound-ing effect of CMB burden The mixed CMB cate-gory has the highest risk of stroke but bydefinition includes only patients with multipleCMBs by contrast the strictly lobar and strictlydeep CMB categories could include patients witha single CMB Specifically we could not fullyinvestigate the independent risk associated with

strictly lobar CMBs critical to the diagnosis ofCAA4 with high recurrent ICH risk1930 Althoughwe undertook meta-regression this can only par-tially account for confounding and is unlikely tofully account for variables such as age and hyper-tension Our logistic regression assumes an averageCMB count within each category (difficult to esti-mate in the open-ended $10 CMB category) andthat the log odds of ICHIS increase linearly withCMB burden However consistent findings from 2complementary statistical analyses strongly supportthe hypothesis that increasing CMB burden in-creases the risk of ICH more than that of IS Fur-ther limitations include the variable study samplesize and follow-up which may bias our resultsespecially regarding ICH a rare outcome with wideCIs around risk estimates (which overlap for thedifferent CMB burden categories) A time to eventanalysis may have been a more appropriate statisti-cal method given the varying follow-up but thiswas not possible with the data available Imagingprotocols and analysis were similar but not com-pletely uniform (table e-2) field strength31 echotime32 and optimized paramagnetic sequences33

can all influence CMB detection However moststudies were performed at 15T with echo timeswithin a narrow range making this unlikely toaffect our conclusions Nevertheless our resultsare only generalizable to patients scanned on15T MRI using gradient recalled echo and maynot be applicable to susceptibility-weighted imag-ing (SWI) or MRI with a higher field strength SWIincreases the number of CMBs detected3334 com-pared to T2-weighted images CMB burden cate-gories may thus have to be revised for SWI Finallystudies used different methods for CMB ratingstandardized rating instruments13536 may improvethe reliability of defining CMB categories partic-ularly that of a single CMB

Although our study provides important newinformation to fully determine how CMBs mightinfluence antithrombotic decisions the interactionbetween CMBs and antiplatelet agents and antico-agulants needs to be further addressed in large pro-spective studies Although the prevalence ofantithrombotic and anticoagulant use did notshow an association with either ICH or IS outcomein our meta-regression the large majority of pa-tients we included were treated with antiplateletagents Very few patients included in our meta-analysis were on anticoagulation more data aretherefore needed on this group who may be athighest ICH risk Ongoing prospective observa-tional studies addressing this question includeuclacukcromis-237 and clinicaltrialsgovct2showNCT02238470 Further pooled analyses of



individual patient data from these and other obser-vational studiesmdashand ultimately randomizedcontrolled trials based on CMB burdenmdashareneeded to fully assess the interaction betweenCMBs and antithrombotic drugs (both antiplateletagents and anticoagulants) after IS and TIA Nev-ertheless we have shown that with increasingCMB burden the risk of ICH increases moresteeply than that of IS in a cohort of IS and TIApatients largely treated with antiplatelet medica-tion A high CMB burden (eg $5 CMBs) mayidentify patients at similar or greater risk of ICHthan IS with implications for antithrombotictreatment and future randomized controlled trials

AUTHOR CONTRIBUTIONSDW was involved in data acquisition and statistical analysis and drafted

the manuscript AC was involved in study concept data acquisition and

revised the manuscript GA was involved in statistical analysis

and revised the manuscript ZVF was involved in statistical analysis

and revised the manuscript PR was involved in data acquisition and

revised the manuscript SG was involved in data acquisition and revised

the manuscript TI contributed data was involved in data acquisition

and revised the manuscript FF contributed data was involved in data

acquisition and revised the manuscript HN contributed data was

involved in data acquisition and revised the manuscript RV contrib-

uted data was involved in data acquisition and revised the manuscript

PMR contributed data was involved in data acquisition and revised

the manuscript VIHK contributed data was involved in data acquisi-

tion and revised the manuscript VT contributed data was involved in

data acquisition and revised the manuscript Y-SL contributed data

was involved in data acquisition and revised the manuscript YDK con-

tributed data was involved in data acquisition and revised the manu-

script YH contributed data was involved in data acquisition and

revised the manuscript KSW contributed data was involved in data

acquisition and revised the manuscript HRJ was involved in study

concept and revised the manuscript DJW was involved in study con-

cept contributed data and revised the manuscript

STUDY FUNDINGDr Werring receives research support from the Stroke Association Brit-

ish Heart Foundation and the Rosetrees Trust Dr Wilson receives

research support from the Stroke Association British Heart Foundation

This work was undertaken at UCLHUCL which received a proportion

of funding from the Department of Healthrsquos NIHR Biomedical Research

Centres funding scheme Dr Rothwell receives research support from the

Wellcome Trust Wolfson Foundation UK Stroke Association British

Heart Foundation National Institute of Health Research (NIHR) Med-

ical Research Council Dunhill Medical Trust and the NIHR Biomedical

Research Centre Oxford Dr Rothwell is also in receipt of an NIHR

Senior Investigator Award and a Wellcome Trust Senior Investigator

Award Dr Gareth Ambler receives funding from the National Institute

for Health Research University College London Hospitals Biomedical

Research Centre

DISCLOSURED Wilson A Charidimou G Ambler Z Fox S Gregoire P Rayson

T Imaizumi F Fluri H Naka S Horstmann R Veltkamp P Roth-

well and V Kwa report no disclosures relevant to the manuscript

V Thijs was affiliated to KU Leuven and funded by FWO Flanders

and VIB Flanders when this research was performed Y Lee Y Kim

Y Huang K Wong H Jaumlger and D Werring report no disclosures

relevant to the manuscript Go to Neurologyorg for full disclosures

Received March 9 2016 Accepted in final form June 21 2016

REFERENCES1 Greenberg SM Vernooij MW Cordonnier C et al Cere-

bral microbleeds a guide to detection and interpretation

Lancet Neurol 20098165ndash174

2 Fazekas F Kleinert R Roob G et al Histopathologic

analysis of foci of signal loss on gradient-echo T2-

weighted MR images in patients with spontaneous

intracerebral hemorrhage evidence of microangiopathy-

related microbleeds AJNR Am J Neuroradiol 199920

637ndash642

3 Tanaka A Ueno Y Nakayama Y Takano K Takebayashi S

Small chronic hemorrhages and ischemic lesions in associa-

tion with spontaneous intracerebral hematomas Stroke

1999301637ndash1642

4 Knudsen KA Rosand J Karluk D Greenberg SM Clin-

ical diagnosis of cerebral amyloid angiopathy validation of

the Boston criteria Neurology 200156537ndash539

5 Shoamanesh A Kwok CS Benavente O Cerebral micro-

bleeds histopathological correlation of neuroimaging

Cerebrovasc Dis 201132528ndash534

6 Charidimou A Krishnan A Werring DJ Rolf Jager H

Cerebral microbleeds a guide to detection and clinical rel-

evance in different disease settings Neuroradiology 2013

55655ndash674

7 Imaizumi T Inamura S Kohama I Yoshifuji K

Nomura T Komatsu K Nascent lobar microbleeds

and stroke recurrences J Stroke Cerebrovas Dis 2014

23610ndash617

8 Martinez-Ramirez S Romero JR Shoamanesh A et al

Diagnostic value of lobar microbleeds in individuals without

intracerebral hemorrhage Alzheimerrsquos Dement 201511

1480ndash1488

9 Soo YO Yang SR Lam WW et al Risk vs benefit of anti-

thrombotic therapy in ischaemic stroke patients with cere-

bral microbleeds J Neurol 20082551679ndash1686

10 Lovelock CE Cordonnier C Naka H et al Antithrom-

botic drug use cerebral microbleeds and intracerebral

hemorrhage a systematic review of published and unpub-

lished studies Stroke 2010411222ndash1228

11 Boulanger JM Coutts SB Eliasziw M et al Cerebral

microhemorrhages predict new disabling or fatal strokes

in patients with acute ischemic stroke or transient ischemic

attack Stroke 200637911ndash914

12 Thijs V Lemmens R Schoofs C et al Microbleeds and

the risk of recurrent stroke Stroke 2010412005ndash2009

13 Lim JS Hong KS Kim GM et al Cerebral microbleeds

and early recurrent stroke after transient ischemic attack

results from the Korean Transient Ischemic Attack Expres-

sion Registry JAMA Neurol 201572301ndash308

14 Fluri F Jax F Amort M et al Significance of microbleeds

in patients with transient ischaemic attack Eur J Neurol

201219522ndash524

15 Janaway BM Simpson JE Hoggard N et al Brain haemo-

siderin in older people pathological evidence for an ischaemic

origin of magnetic resonance imaging (MRI) microbleeds

Neuropathol Appl Neurobiol 201440258ndash269

16 Grutzendler J Murikinati S Hiner B et al Angiophagy

prevents early embolus washout but recanalizes microvessels

through embolus extravasation Sci Transl Med 20146

226ra231

17 Tanskanen M Makela M Myllykangas L Rastas S

Sulkava R Paetau A Intracerebral hemorrhage in the oldest

old a population-based study (VANTAA 851) Front Neurol

20123103



18 Akoudad S Ikram MA Koudstaal PJ et al Cerebral mi-

crobleeds are associated with the progression of ischemic

vascular lesions Cerebrovasc Dis 201437382ndash388

19 Greenberg SM Eng JA Ning M Smith EE Rosand J

Hemorrhage burden predicts recurrent intracerebral hemor-

rhage after lobar hemorrhage Stroke 2004351415ndash1420

20 Charidimou A Kakar P Fox Z Werring DJ Cerebral

microbleeds and recurrent stroke risk systematic review

and meta-analysis of prospective ischemic stroke and tran-

sient ischemic attack cohorts Stroke 201344995ndash1001

21 Huang Y Cheng Y Wu J et al Cilostazol as an alternative

to aspirin after ischaemic stroke a randomised double-

blind pilot study Lancet Neurol 20087494ndash499

22 Kwa VI Algra A Brundel M Bouvy W Kappelle LJ

Group MS Microbleeds as a predictor of intracerebral

haemorrhage and ischaemic stroke after a TIA or minor

ischaemic stroke a cohort study BMJ Open 20133

e002575

23 Naka H Nomura E Takahashi T et al Combinations of

the presence or absence of cerebral microbleeds and

advanced white matter hyperintensity as predictors of sub-

sequent stroke types AJNR Am J Neuroradiol 200627

830ndash835

24 Song TJ Kim J Lee HS et al The frequency of cerebral

microbleeds increases with CHADS(2) scores in stroke

patients with non-valvular atrial fibrillation Eur J Neurol

201320502ndash508

25 Fan YH Zhang L Lam WW Mok VC Wong KS Cere-

bral microbleeds as a risk factor for subsequent intracere-

bral hemorrhages among patients with acute ischemic

stroke Stroke 2003342459ndash2462

26 Mok VC Lau AY Wong A et al Long-term prognosis of

Chinese patients with a lacunar infarct associated with

small vessel disease a five-year longitudinal study Int J

Stroke 2009481ndash88

27 Imaizumi T Horita Y Hashimoto Y Niwa J Dotlike

hemosiderin spots on T2-weighted magnetic resonance

imaging as a predictor of stroke recurrence a prospective

study J Neurosurg 2004101915ndash920

28 Wang DN Hou XW Yang BW Lin Y Shi JP Wang N

Quantity of cerebral microbleeds antiplatelet therapy and

intracerebral hemorrhage outcomes a systematic review

and meta-analysis J Stroke Cerebrovasc Dis 201524

2728ndash2737

29 Antithrombotic Trialists C Collaborative meta-analysis of

randomised trials of antiplatelet therapy for prevention of

death myocardial infarction and stroke in high risk pa-

tients BMJ 200232471ndash86

30 Akoudad S Portegies ML Koudstaal PJ et al Cerebral

microbleeds are associated with an increased risk of stroke

the Rotterdam study Circulation 2015132509ndash516

31 Stehling C Wersching H Kloska SP et al Detection of

asymptomatic cerebral microbleeds a comparative study at

15 and 30 T Acad Radiol 200815895ndash900

32 Gregoire SM Werring DJ Chaudhary UJ et al Choice of

echo time on GRE T2-weighted MRI influences the

classification of brain microbleeds Clin Radiol 201065

391ndash394

33 Cheng AL Batool S McCreary CR et al Susceptibility-

weighted imaging is more reliable than T2-weighted

gradient-recalled echo MRI for detecting microbleeds


34 Shams S Martola J Cavallin L et al SWI or T2 which

MRI sequence to use in the detection of cerebral micro-

bleeds The Karolinska Imaging Dementia Study AJNR

Am J Neuroradiol 2015361089ndash1095

35 Cordonnier C Al-Shahi Salman R Wardlaw J Spontane-

ous brain microbleeds systematic review subgroup analy-

ses and standards for study design and reporting Brain

20071301988ndash2003

36 Gregoire SM Chaudhary UJ Brown MM et al The

Microbleed Anatomical Rating Scale (MARS) reliability

of a tool to map brain microbleeds Neurology 200973

1759ndash1766

37 Charidimou A Wilson D Shakeshaft C et al The Clinical

Relevance of Microbleeds in Stroke study (CROMIS-2)

rationale design and methods Int J Stroke 201510(Suppl

A100)155ndash161

Neurologyorg Offers Important Information toPatients and Their Families

The Neurologyreg Patient Page provides

bull A critical review of ground-breaking discoveries in neurologic research that are written especiallyfor patients and their families

bull Up-to-date patient information about many neurologic diseases

bull Links to additional information resources for neurologic patients

AllNeurology Patient Page articles can be easily downloaded and printed and may be reproduced todistribute for educational purposes Click on the lsquoPatientsrsquo link on the home page (Neurologyorg) for acomplete index of Patient Pages



DOI 101212WNL00000000000031832016871501-1510 Published Online before print September 2 2016Neurology

Duncan Wilson Andreas Charidimou Gareth Ambler et al meta-analysis

Recurrent stroke risk and cerebral microbleed burden in ischemic stroke and TIA A

This information is current as of September 2 2016

rights reserved Print ISSN 0028-3878 Online ISSN 1526-632X1951 it is now a weekly with 48 issues per year Copyright copy 2016 American Academy of Neurology All

reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology

ServicesUpdated Information amp

httpnneurologyorgcontent87141501fullincluding high resolution figures can be found at

Supplementary Material

183DC1httpnneurologyorgcontentsuppl20160902WNL0000000000003Supplementary material can be found at

References httpnneurologyorgcontent87141501fullref-list-1

This article cites 37 articles 16 of which you can access for free at

Citations httpnneurologyorgcontent87141501fullotherarticles

This article has been cited by 4 HighWire-hosted articles

Subspecialty Collections

httpnneurologyorgcgicollectionmriMRI

httpnneurologyorgcgicollectionintracerebral_hemorrhageIntracerebral hemorrhage

httpnneurologyorgcgicollectioninfarctionInfarction e

httpnneurologyorgcgicollectionall_cerebrovascular_disease_strokAll Cerebrovascular diseaseStrokefollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online



a strictly lobar distribution are highly specificfor cerebral amyloid angiopathy (CAA) whichcauses progressive deposition of b-amyloid insmall cortical and leptomeningeal arterialwalls6 though this pattern may not be so spe-cific in Eastern (Asian) people7 and in thosewithout ICH8

Multiple prospective studies in ischemicstroke (IS) cohorts have shown that CMBsare associated with subsequent ICH risk910

However CMBs are also associated withincreased subsequent IS risk11ndash14 Indeed sug-gested ischemic mechanisms for CMBsinclude ischemia-mediated iron store releaseby oligodendrocytes15 phagocytosis of red cellmicroemboli into the perivascular space(termed angiophagy)16 or hemorrhagic trans-formation of small microinfarcts17 Indeed ina recent community study after adjusting forcardiovascular risk factors CMBs were foundto be associated with lacunes and white mattervolume progression18 Few data are availableon how CMB burden affects the balance ofICH and IS risk in different populations InCAA cohorts an increasing number of CMBsis associated with an increased risk of ICHsuggesting a relationship between CMB num-ber and the severity of bleeding-prone arterio-pathy19 Whether an increasing number ofCMBs is also associated with an increased riskof ICH in IS and TIA cohorts remains uncer-tain If increasing CMB burden shifts the bal-ance of risk toward ICH rather than IS thiscould have major clinical relevance for antith-rombotic risk-benefit decisions after IS andTIA Our previous meta-analysis of 10 pro-spective studies including 3067 patients withIS or TIA found that CMB presence is associ-ated with a higher risk of ICH than IS20 (theodds ratio was 853 for ICH and 155 for IS)but was not able to address the key clinicalquestion of how the number (burden) ofCMBs influences ICH and IS risk

We therefore performed a pooled analysis ofaggregate summary data including CMB bur-den and distribution to investigate the risk ofsubsequent IS and ICH in individuals who havehad an IS or TIA We tested the followinghypotheses (1) CMB presence is associatedwith an increased risk of stroke (ICH IS)and (2) as CMB burden increases (due to amore

severe bleeding-prone arteriopathy) the risk ofICH increases more steeply than the risk of IS

METHODS We searched Medline and Embase from 1996 (the

year CMBs were first reported) through to April 2015 Our

search strategy was as follows

1 ldquoCerebral microbleedrdquo or CMB or ldquocerebral microhemorrrdquo

or ldquobrain microbleedrdquo or ldquobrain microhemorrrdquo

2 Stroke or ldquoischemic strokerdquo or TIA or ldquointrac adj2 h

emorrhagrdquo or ICH

3 1 and 2

We included published and unpublished studies fulfilling the

following criteria (1) performed paramagnetic-sensitive MRI se-

quences to detect CMBs at baseline (2) assessed CMBs at base-

line and associations with IS or ICH as primary or secondary

outcomes (3) had a prospective study design with at least 3

months of follow-up and (4) fulfilled at least 4 of 6 predefined

quality indicators We excluded cross-sectional studies and case

series Two clinical research fellows (AC and DW) reviewed

each study for eligibility

Data extraction We contacted all authors to provide data

on study population size patient-year follow-up and

antithrombotic treatment We obtained data on outcome

events of symptomatic IS and ICH with baseline CMB number

categories as follows CMB present 1 CMB 1 CMB 2ndash4

CMBs 5ndash10 CMBs 10 CMBs strictly deep CMBs strictly

lobar CMBs and mixed distribution CMBs We extracted

all demographic imaging and follow-up outcome data

from each study

Quality assessment and reducing the risk of bias All

included studies were critically appraised against a checklist of 6

key quality indicators (table e-1 at Neurologyorg) with

reference to the Strengthening the Reporting of Observational

Studies in Epidemiology statement and the Preferred Reporting

Items for Systematic Reviews and Meta-Analyses guidelines The

quality criteria included assessment for bias and all studies had

a quality score of $46

Statistical methods We first performed separate random effect

meta-analyses to derive summary estimates of the pooled risk

ratios of ICH and IS for each CMB category vs the reference

category of ldquono CMBsrdquo Due to the small number of events in

some studies the categories 5ndash10 and 10 were combined in

line with previous studies of CMB burden and prognosis which

demonstrated their prognostic relevance for future ICH risk19

Logistic regression was then used to estimate the increased risk

(odds) of ISICH for each additional CMB First CMB

categories were converted to a continuous scale by assuming

that patients in the CMB groups 0 1 2ndash4 5ndash10 and 10

had on average 0 1 3 75 and 125 CMBs respectively

Then for each study a logistic regression model was fitted

relating the (log) odds of IS or ICH to the (estimated) number

of CMBs These (log) odds ratios were then pooled using random

effects meta-analysis

We calculated the I2 statistic to investigate heterogeneity

Funnel plots (Begg and Mazumdar) were generated to investigate

publication bias Finally where necessary we undertook meta-

regression of confounding covariates of biological plausibility or

with differences between the studies (average follow-up age

hypertension prevalence demographics antithrombotic use)

All analyses were performed using STATA 120 (StataCorp LP

College Station TX)




















CMBdistribution n




PooledRR

Lower95 CI

Upper95 CI



PooledRR

Lower95 CI

Upper95 CI

CMB presence 1151284 (9) 34 18 14 25 491142 (43) 38 63 35 114

1 CMB 31433 (72) 18 18 10 31 8354 (23) 17 46 19 107

2ndash4 CMBs 44433 (102) 48 24 13 44 9383 (23) 18 56 24 133

Dagger5 CMBs 34342 (105) 51 27 15 49 24274 (88) 82 141 69 290

Strictly lobar 31332 (93) 39 20 14 29 12332 (36) 32 105 45 243

Strictly deep 29437 (66) 12 16 10 27 6437 (14) 1 33 13 85

Mixed 44411 (107) 53 26 15 43 25411 (61) 57 111 55 226
































































Research Centre

















637ndash642




1999301637ndash1642










55655ndash674




23610ndash617




1480ndash1488




















201219522ndash524








226ra231




20123103




















e002575





830ndash835




201320502ndash508

















2728ndash2737














391ndash394












20071301988ndash2003




1759ndash1766




A100)155ndash161






























Reprints





















CMBdistribution n




PooledRR

Lower95 CI

Upper95 CI



PooledRR

Lower95 CI

Upper95 CI

CMB presence 1151284 (9) 34 18 14 25 491142 (43) 38 63 35 114

1 CMB 31433 (72) 18 18 10 31 8354 (23) 17 46 19 107

2ndash4 CMBs 44433 (102) 48 24 13 44 9383 (23) 18 56 24 133

Dagger5 CMBs 34342 (105) 51 27 15 49 24274 (88) 82 141 69 290

Strictly lobar 31332 (93) 39 20 14 29 12332 (36) 32 105 45 243

Strictly deep 29437 (66) 12 16 10 27 6437 (14) 1 33 13 85

Mixed 44411 (107) 53 26 15 43 25411 (61) 57 111 55 226
































































Research Centre

















637ndash642




1999301637ndash1642










55655ndash674




23610ndash617




1480ndash1488




















201219522ndash524








226ra231




20123103




















e002575





830ndash835




201320502ndash508

















2728ndash2737














391ndash394












20071301988ndash2003




1759ndash1766




A100)155ndash161






























Reprints
































































Research Centre

















637ndash642




1999301637ndash1642










55655ndash674




23610ndash617




1480ndash1488




















201219522ndash524








226ra231




20123103




















e002575





830ndash835




201320502ndash508

















2728ndash2737














391ndash394












20071301988ndash2003




1759ndash1766




A100)155ndash161






























Reprints





















e002575





830ndash835




201320502ndash508

















2728ndash2737














391ndash394












20071301988ndash2003




1759ndash1766




A100)155ndash161






























Reprints

























Reprints




Recurrentstrokeriskandcerebralmicrobleed burden in ...burden in ischemic stroke and TIA A...

Documents

Transcript of Recurrentstrokeriskandcerebralmicrobleed burden in ...burden in ischemic stroke and TIA A...