Extensive Leukoaraiosis Is Associated With High Early
Risk of Recurrence After Ischemic Stroke
Gyeong-Moon Kim, MD, PhD; Kwang-Yeol Park, MD; Ross Avery, BS;
Johanna Helenius, MD; Natalia Rost, MD; Jonathan Rosand, MD, MSc;
Bruce Rosen, MD, PhD; Hakan Ay, MD
Downloaded from http://stroke.ahajournals.org/ by guest on June 15, 2017
Background and Purpose—The integrity of white matter tracts connecting different parts of the brain is important for
rapid compensation for the lost function from ischemic stroke. Impaired white matter reserve capacity secondary to
leukoaraiosis may facilitate detection of new symptomatic ischemic events that would otherwise remain inconspicuous
after an initial ischemic stroke. We sought to identify whether the extent of leukoaraiosis was a predictor of risk of early
stroke recurrence.
Methods—We used Cox regression analysis in consecutive patients with ischemic stroke to determine the relationship
between leukoaraiosis burden and symptomatic stroke recurrence within 90 days. We graded total leukoaraiosis,
periventricular leukoaraiosis, and subcortical leukoaraiosis using the Fazekas scale as mild (<2) and extensive (≥2) on
fluid-attenuated inversion recovery images obtained within 72 hours of stroke onset in the hemisphere contralateral to
acute stroke.
Results—There were 106 recurrent events in 2378 patients. The cumulative incidence of recurrence was 5.9% at 90 days.
Kaplan–Meier estimate of recurrence-free survival rate was lower in patients with extensive leukoaraiosis (P=0.04) and
extensive periventricular leukoaraiosis (P=0.02) but not in extensive subcortical leukoaraiosis (P=0.09). Multivariable
Cox regression analysis revealed a hazard ratio of 1.50 (95% confidence interval, 1.00–2.25) for extensive leukoaraiosis,
1.67 (95% confidence interval, 1.11–2.51) for extensive periventricular leukoaraiosis, and 1.42 (95% confidence interval,
0.94–2.12) for extensive subcortical leukoaraiosis.
Conclusions—The extent of leukoaraiosis independently predicts 90-day recurrent stroke risk after ischemic stroke. This
suggests that leukoaraiosis may be used for risk stratification in ischemic stroke. (Stroke. 2014;45:479-485.)
Key Words: leukoaraiosis ◼ magnetic resonance imaging ◼ stroke
T
he preserved integrity of white matter tracts connecting
different parts of the brain contributes to the speed and
the magnitude of compensation for the lost function after ischemic stroke.1,2 An earlier study has shown that, in an event
of acute brain infarction, the odds of developing severe and
persistent symptoms when compared with mild and transient
symptoms increases with increasing volume of leukoaraiosis.2
A shift toward more severe strokes by increasing leukoaraiosis severity suggests that acute infarcts otherwise destined
to remain inconspicuous would be more clinically apparent in
the presence of leukoaraiosis. In other words, the presence of
leukoaraiosis may facilitate detection of new ischemic events
occurring after an initial ischemic stroke by making them
more clinically noticeable. In the present study, we tested the
hypothesis that extensive leukoaraiosis at the time of stroke
conferred an increased risk for early ischemic stroke recurrence. If proven, a relationship between leukoaraiosis and early
risk of recurrent ischemic stroke could have both biological
and clinical implications, offering insight into the mechanisms
of recurrent stroke and serving to clarify ­risk-stratification in
individual patients.
Methods
Study Population and Data Collection
We retrospectively analyzed demographic, clinical, laboratory, and
radiographic data collected from consecutive patients admitted to the
Massachusetts General Hospital for acute ischemic stroke between
March 2003 and April 2011. Because most recurrent stroke events
occur early after an initial ischemic stroke, we excluded patients who
were admitted >72 hours of symptom. We also excluded patients who
Received July 26, 2013; final revision received November 12, 2013; accepted November 15, 2013.
From the Department of Neurology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea (G.-M.K.); Department of
Radiology, A.A. Martinos Center for Biomedical Imaging (G.-M.K., K.-Y.P., R.A., J.H., B.R., H.A.) and Department of Neurology, Stroke Service (N.R.,
J.R., H.A.), Massachusetts General Hospital, Harvard Medical School, Boston; and Department of Neurology, Chung-Ang University Hospital, ­ChungAng University College of Medicine, Seoul, Korea (K.-Y.P.).
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.
113.003004/-/DC1.
Correspondence to Hakan Ay, MD, Department of Radiology and Stroke Service, A.A. Martinos Center for Biomedical Imaging, Massachusetts General
Hospital, Harvard Medical School, 149 13th St, Room 2301, Charlestown, MA 02129. E-mail [email protected]
© 2013 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org
DOI: 10.1161/STROKEAHA.113.003004
479
480 Stroke February 2014
Downloaded from http://stroke.ahajournals.org/ by guest on June 15, 2017
did not undergo an MRI study and patients who did not exhibit an
acute infarct on diffusion-weighted MRI (DWI). The study was approved by the local Human Studies Committee.
We collected data on published predictors of short-term stroke recurrence from an institutional database maintained by stroke-trained
physicians. These predictors included age, sex, history of stroke, or
transient ischemic attack within a month preceding the index stroke,
admission National Institute of Health Stroke Scale score, thrombolytic treatment, carotid or cardiac intervention, and pathogenic stroke
subtype determined using the Causative Classification of Stroke
(CCS) system.3 We also collected data on vascular risk factors that
are known to confer increased stroke risk over the long term. These
risk factors included hypertension (blood pressure, ≥140/90 mm Hg
on repeated measurements or prior use of antihypertensive medication), diabetes mellitus (fasting blood glucose level, ≥126 mg/dL
on repeated measurements or the use of medications to lower blood
glucose), atrial fibrillation, coronary artery disease, current smoking,
and distant history of prior stroke (>1 month). Time of index stroke
was determined based on careful interview with patients and reliable
observers. In patients with uncertain onset of stroke, the time at which
the patient was last known to be free of the signs and symptoms of the
stroke was considered to be the time of onset.
Image Acquisition and Analysis
MRI was performed within 24 hours of admission by 1.5T GE
Signa (GE Medical Systems, Milwaukee, WI) or Siemens Sonata
(Siemens Medical Solutions, Erlangen, Germany) scanners. Image
acquisition and processing protocols were previously described in
detail.4 We defined leukoaraiosis as hyperintense lesions on axial T2
­fluid-attenuated inversion recovery (FLAIR) images that are located
in the region starting at the lateral ventricular border and extending
up to the corticomedullary junction. Hyperintense lesions involving
the convolutional white matter, U-fibers, corpus callosum, internal
capsule, and anterior commissure were not regarded as leukoaraiosis.5 The boundaries of leukoaraiosis were differentiated from the
acute ischemic lesion by visually coregistering FLAIR images with
DWI. Discrete territorial brain lesions with well-defined borders that
seemed hyperintense on FLAIR images, such as chronic subcortical
infarcts and lacunar infarcts, were not considered leukoaraiosis. We
graded leukoaraiosis in the hemisphere contralateral to acute stroke
using the Fazekas scale.6 In patients with bihemispheric infarcts, we
performed leukoaraiosis ratings in the hemisphere with lower burden of acute infarcts. On the basis of visual assessment of FLAIR
images, we further categorized leukoaraiosis into periventricular
leukoaraiosis (PLA) and subcortical leukoaraiosis (SLA).6 PLA indicated white matter hyperintensities exclusively located around the
lateral ventricles (0: absent, 1: caps or pencil lining, 2: smooth halo,
3: irregular periventricular hyperintensity extending into deep white
matter). SLA denoted white matter hyperintensities in the centrum
semiovale and the corona radiata (0: absent, 1: punctate foci, 2: beginning confluence of foci 3: large confluent areas). Total leukoaraiosis
score was calculated by summing up the scores for PLA and SLA and
changed between 0 and 6. Large regions of leukoaraiosis that started
from the ventricular border and extended into the deep white matter
were classified as both PLA and SLA. We computed a receiver operating characteristics curve for total leukoaraiosis score and stroke recurrence and identified that the optimal operating point on this curve
corresponded to the score of 2. Based on this, we dichotomized leukoaraiosis according to its severity as mild (Fazekas scores of 0 or
1) and extensive (Fazekas scores, ≥2). Extensive total leukoaraiosis
was defined as having either PLA≥2 or SLA≥2. We tested inter-rater
reliability of dichotomized leukoaraiosis ratings between 2 investigators (J.H. and G.-M.K.) in a set of 100 consecutive patients using κ
statistics.7 The κ coefficient for PLA and SLA were 0.97 and 0.96,
respectively. All leukoaraiosis ratings were done blind to recurrent
stroke status.
Follow-Up Assessment
The outcome variable was recurrent ischemic stroke within 90 days
of index stroke. Recurrent stroke was defined as clinical deterioration
caused by a new infarct that was spatially distinct from the index lesion. All recurrent events were confirmed by brain imaging. Clinical
worsening because of brain edema, hemorrhagic transformation of the
index lesion, and other metabolic and systemic causes were not considered as recurrence. Follow-up data on stroke recurrence were collected
through inspection of inpatient medical record notes and outpatient
assessment notes by investigators who were blinded to leukoaraiosis
information on baseline MRI. All recurrent events were adjudicated by
a separate investigator using the same data sources (H.A.).
Statistical Analyses
Statistical analyses explored relationships between leukoaraiosis
and 90-day stroke recurrence. We made univariate comparisons between patients with and without recurrent stroke using Cox regression analysis. We used independent t tests or the Mann–Whitney U
tests for continuous variables and Pearson χ2 and Fisher exact tests for
categorical variables to compare patients with and without complete
follow-up. We constructed a multivariable Cox proportional hazard
model where time to recurrent stroke was the response and variables
that were associated with leukoaraiosis or stroke recurrence with a
univariate P value <0.1 were covariates. We performed Kaplan–Meier
analysis and log-rank test to assess the predictive value of leukoaraiosis for early risk of stroke recurrence in each pathogenic category.
Results were presented as hazard ratio with 95% confidence interval
(CI). Values of P<0.05 were considered statistically significant. All
statistical analyses were performed using a commercially available
software (SPSS for Windows, version 13.0; SPSS, Chicago, IL).
Results
Study Population and Follow-Up Events
A total of 3269 consecutive patients were admitted with
diagnosis of ischemic stroke during the study period. We
excluded 538 patients who were admitted after 72 hours of
symptom onset. We also excluded 281 patients in whom
MRI was not performed because of contraindications, such
as metal implants, 28 patients in whom MRI was not deemed
to be necessary by the treating physician (often because there
was another MRI obtained at an outside center before coming to the Massachusetts General Hospital), and 44 patients in
whom FLAIR images were not acquired. The remaining 2378
patients comprised the study population.
The median age of the study population was 70 years
(interquartile range, 58–80 years), and the median time from
onset to MRI was 10 hours (interquartile range, 5–21 hours).
Complete 90-day follow-up data were available in 1736
patients (73%). In the remaining 27%, median follow-up was
7 days (interquartile range, 4–21 days). Patients with incomplete follow-up were older and more likely to have CCS subtype of small artery occlusion. The Fazekas score and other
baseline characteristics listed in Table 1 were similar between
patients with and without complete follow-up.
A total of 106 patients developed recurrent ischemic stroke
within 90 days of index ischemic stroke. The cumulative incidence of recurrence by the Kaplan–Meier estimate was 1.8% at 7
days, 2.6% at 14 days, and 5.9% at 90 days. Baseline demographics and clinical characteristics of the cohort are summarized in
Table 1. Patients with recurrent stroke were more likely to have
high blood pressure at admission, prior stroke, or transient ischemic attack within the month preceding the index stroke, chronic
Kim et al Leukoaraiosis and Early Stroke Recurrence 481
Table 1. Demographic and Clinical Characteristics According to Recurrence Status
Recurrence (n=106)
No Recurrence (n=2272)
HR (95% CI)
P Value
68.5 (15.2)
67.9 (15.4)
1.01 (1.00–1.02)
0.253
50 (47.2)
1045 (46.0)
1.07 (0.73–1.57)
0.724
Hypertension
85 (80.2)
1550 (68.2)
1.99 (1.23–3.21)
0.005
Diabetes mellitus
24 (22.6)
524 (23.1)
0.98 (0.62–1.54)
0.923
Atrial fibrillation
23 (21.7)
577 (25.4)
0.94 (0.59–1.49)
0.776
Current smoking
20 (18.9)
437 (19.2)
0.92 (0.56–1.51)
0.739
Prior TIA/stroke <1 mo
36 (34.0)
301 (13.2)
3.00 (2.01–4.48)
<0.001
Presence of chronic
infarcts on MRI
25 (23.6)
335 (14.7)
1.74 (1.11–2.73)
0.015
3 (0–6)
4 (1–10)
0.97 (0.93–1.00)
0.041
Large artery atherosclerosis
46 (43.3)
474 (20.9)
3.96 (1.69–9.26)
Cardioaortic embolism
31 (29.2)
1044 (46.0)
1.35 (0.56–3.23)
Small artery occlusion
2 (1.9)
294 (12.9)
0.30 (0.06–1.48)
21 (19.8)
174 (7.7)
4.84 (1.96–12.0)
6 (5.7)
286 (12.6)
Reference
Antiplatelet
75 (70.8)
1594 (70.2)
0.88 (0.58–1.33)
0.530
Anticoagulation
44 (41.5)
840 (37.0)
1.12 (0.76–1.64)
0.574
Antihypertensive
28 (26.4)
634 (27.9)
0.87 (0.57–1.34)
0.534
Statin
72 (67.9)
1494 (65.8)
0.95 (0.63–1.43)
0.802
Cerebral revascularization
11 (10.4)
117 (5.1)
1.73 (0.93–3.23)
0.085
Age, y, mean (SD)
Women, n (%)
Risk factors, n (%)
NIHSS, median (IQR)
CCS subtype, n (%)
Downloaded from http://stroke.ahajournals.org/ by guest on June 15, 2017
Other uncommon causes
Undetermined
<0.001
Treatment, n (%)
CCS indicates Causative Classification of Stroke; CI, confidence interval; HR, hazard ratio; IQR, interquartile range; NIHSS, National
Institute of Health Stroke Scale; and TIA, transient ischemic attack.
infarcts on MRI, lower admission National Institute of Health
Stroke Scale score, and CCS subtypes of large artery atherosclerosis and other uncommon causes (P<0.001). The category of
other uncommon causes included acute arterial dissection (52%),
hypercoagulability disorders (18%), and other rare conditions,
such as vasculitis, thrombosed aneurysm, radiation, etc. (30%).
recent prior transient ischemic attack/stroke, chronic infarcts,
and CCS subtype revealed a hazard ratio of 1.50 (95% CI, 1.00–
2.25) for extensive leukoaraiosis, 1.67 (95% CI, 1.11–2.51) for
extensive PLA, 1.42 (95% CI, 0.94–2.12) for extensive SLA,
and 1.63 (95% CI, 1.07–2.48) for both extensive PLA and
extensive SLA (Table 3).
Leukoaraiosis and Stroke Recurrence
Leukoaraiosis and Pathogenic Stroke Subtypes
Table I in the online-only Data Supplement shows recurrent
stroke rates by Fazekas score. There was extensive leukoaraiosis
in 1056 (44.4%), extensive PLA in 954 (40.2%), extensive SLA
in 747 (31.2%), and both extensive PLA and extensive SLA in
645 (27.1%) patients (Table 2). Recurrent stroke occurred in 56
patients with extensive leukoaraiosis, 53 patients with extensive PLA, and 39 patients with extensive SLA. The probability
of recurrence increased by increasing total Fazekas score (logrank test, P=0.034). Patients with recurrence had marginally
higher rate of extensive leukoaraiosis (P=0.074) and significantly higher rate of extensive PLA (P=0.034) when compared
with the patients without recurrence. There was no difference
in extensive SLA (P=0.222) between patients with and without
recurrence. Kaplan–Meier estimate of recurrence-free survival
rate was lower in patients with extensive leukoaraiosis and PLA,
whereas there was only a marginal reduction in recurrence-free
survival rate in extensive SLA (Figure). Multivariable Cox proportional hazards regression models that adjusted for hypertension, admission National Institute of Health Stroke Scale score,
Because PLA attained higher statistical significance in prior
analyses, we further explored associations between PLA
Table 2. Leukoaraiosis and 90-Day Stroke Recurrence
Recurrence (%),
(n=106)
No Recurrence (%),
(n=2272)
P Value
Negative (n=1322)
50 (47.2)
1272 (56.0)
0.074
Positive (n=1056)
56 (52.8)
1000 (44.0)
Negative (n=1424)
53 (50.0)
1371 (60.3)
Positive (n=954)
53 (50.0)
901 (39.7)
Negative (n=1631)
67 (63.2)
1564 (68.8)
Positive (n=747)
39 (36.8)
708 (31.0)
Leukoaraiosis
Extensive LA
Extensive PLA
0.034
Extensive SLA
PLA indicates
leukoaraiosis.
periventricular
leukoaraiosis;
and
SLA,
0.222
subcortical
482 Stroke February 2014
the risk of recurrence was lowest (1.0%), whereas it was less
prevalent in the category of other uncommon causes (18%)
where the risk was highest (12.7%). To characterize CCS
subtype–leukoaraiosis–recurrence relationship further, we
calculated hazard ratio for 90-day recurrence by extensive
PLA in each CCS category. The effect of extensive PLA on
stroke recurrence tended to be higher in CCS subtypes with
greater baseline risk (P=0.1); there was more than 2-fold difference in hazard ratio by PLA between the highest and the
lowest risk CCS subtypes (Table 4).
Discussion
Downloaded from http://stroke.ahajournals.org/ by guest on June 15, 2017
Figure. Kaplan–Meier recurrence-free survival curve by extensive
leukoaraiosis (LA). Kaplan–Meier survival curves demonstrating a
reduction in recurrence-free survival by extensive (Fazekas score,
≥2) LA (A) and extensive periventricular LA (PLA; B), and a trend
toward reduction with extensive subcortical LA (SLA; C).
and pathogenic subtype of index stroke. The 90-day cumulative stroke rate by the Kaplan–Meier estimate varied by
pathogenic CCS subtype (log-rank test, P<0.001; Table 4).
Similarly, the prevalence of extensive PLA also varied by
CCS subtype (log-rank test, P<0.001). There was disconcordance between the baseline risk by CCS subtype and the
prevalence of extensive PLA. For instance, extensive PLA
was more prevalent in small artery occlusion (51%) where
Increasing leukoaraiosis burden, whether measured among a
cross section of healthy adults, or measured among patients
with ischemic stroke, has substantial negative consequences
on brain function. Leukoaraiosis is associated with progression of cognitive impairment, gait abnormalities, poor functional outcome, increased mortality, higher risk of hemorrhage
after thrombolysis, and long-term recurrence after ischemic
stroke.8–13 This study extends our knowledge on the relationship between leukoaraiosis and ischemic stroke by showing
that leukoaraiosis is also a risk factor for early stroke recurrence. After controlling for initial stroke severity, conventional
stroke risk factors, underlying stroke mechanism, and preventive stroke treatment, patients with extensive leukoaraiosis
were 1.5× more likely to have another ischemic stroke in the
short-term than those without extensive leukoaraiosis.
Extensive leukoaraiosis in the periventricular location was
associated with slightly more prominent increase in risk of
early recurrence when compared with subcortical location.
This finding is consistent with prior observations in other
conditions; the extent of PLA when compared with SLA
more strongly correlates with cognitive decline in the nondemented elderly population,14–16 progression from mild cognitive impairment to Alzheimer disease,17 burden of aortic
atherosclerosis,18 poor functional outcome after stroke,11 and
risk of recurrent hemorrhage in patients with lobar intracerebral hemorrhage.19 It has been suggested that SLA predominantly disrupts the short association fibers that connect
adjacent gyri to each other, whereas PLA affects the long
association fibers that connect multiple distant cortical areas
and is, therefore, more likely to interfere with cognitive and
executive functions.20 Alternatively, SLA and PLA may be
representing pathophysiologically different states with different burden of injury in the white matter. In support of this,
recent evidence suggests that PLA is primarily associated
with diminished cerebral vasomotor reactivity and subsequent cerebral hypoperfusion,21 whereas SLA is associated
with microangiopathy.22
The exact mechanism by which leukoaraiosis confers
increased risk of short-term stroke recurrence is not known.
The severity of leukoaraiosis is, in part, influenced by vascular risk factors, such as hypertension, advanced age, diabetes
mellitus, and smoking. However, the association between leukoaraiosis and recurrent stroke risk cannot be explained solely
on the basis of shared risk factors because the present study
and previously published studies clearly indicate that leukoaraiosis is a predictor of stroke risk even after controlling for
Kim et al Leukoaraiosis and Early Stroke Recurrence 483
Table 3. Predictors of Stroke Recurrence Within 90-Days: Results of the Cox Proportional
Hazard Regression Analysis
Model 1 (Extensive LA)
HR (95% CI)
P Value
Model 2 (Extensive PLA)
Model 3 (Extensive SLA)
HR (95% CI)
HR (95% CI)
P Value
P Value
Hypertension
2.12 (1.26–3.56)
0.005
2.09 (1.25–3.52)
0.005
2.20 (1.32–3.68)
0.003
NIHSS
0.97 (0.93–1.00)
0.065
0.97 (0.94–1.00)
0.069
0.97 (0.93–1.00)
0.059
Previous TIA or
stroke <1 mo
2.13 (1.37–3.30)
<0.001
2.16 (1.40–3.34)
0.001
2.10 (1.36–3.25)
0.001
Chronic infarcts
on MRI
1.43 (0.90–2.29)
0.130
1.39 (0.87–2.22)
0.17
1.48 (0.93–2.36)
0.095
Extensive LA
1.50 (1.00–2.25)
0.049
...
...
...
...
Extensive PLA
...
...
1.67 (1.11–2.51)
0.013
...
...
Extensive SLA
...
...
...
...
1.42 (0.94–2.13)
0.095
Large artery
atherosclerosis
3.25 (1.38–7.67)
0.007
3.20 (1.35–7.55)
0.008
3.28 (1.39–7.75)
0.007
Cardioaortic
embolism
1.53 (0.64–3.67)
0.344
1.48 (0.62–3.57)
0.377
1.53 (0.64–3.70)
0.336
Small artery
occlusion
0.26 (0.05–1.31)
0.103
0.26 (0.05–1.27)
0.095
0.27 (0.06–1.34)
0.110
Other uncommon
causes
7.28 (2.89–18.4)
<0.001
7.35 (2.91–18.6)
<0.001
7.14 (2.83–18.0)
<0.001
Reference
...
Reference
...
Reference
...
CCS subtype
Downloaded from http://stroke.ahajournals.org/ by guest on June 15, 2017
Undetermined
CCS indicates Causative Classification of Stroke; CI, confidence interval; HR, hazard ratio; PLA, periventricular leukoaraiosis;
and SLA, subcortical leukoaraiosis.
vascular risk factors.2,13,15,23 Of note, none of the conventional
factors except for hypertension that confer an increased risk in
the long-term predicted stroke recurrence in the short-term in
the present study, consistent with the notion that leukoaraiosis
influences short-term risk of recurrent stroke through other
mechanisms. Prior studies have suggested that the extent of
leukoaraiosis can negatively affect the brain’s capacity to tolerate an ischemic insult; among patients with cerebral ischemia,
those with extensive leukoaraiosis are more likely to develop
completed cerebral infarction,6 and among patients with acute
infarction, those with extensive leukoaraiosis are more likely
to develop severe and persistent clinical deficits.1,2 Given that
asymptomatic new infarcts are substantially more common, as
Table 4. Baseline Risk, Prevalence of PLA, and Hazard Ratio
of PLA in Each Pathogenic Category
Estimated 90-D
Recurrence
Rate, %
Prevalence
of Extensive
PLA, %
HR by Extensive
PLA for Early
Recurrence
(95% CI)
11.4
41
2.22 (1.23–3.99)
Cardioaortic
embolism
3.8
42
1.88 (0.93–3.83)
Small artery
occlusion
1.0
51
0.87 (0.05–13.8)
12.7
18
2.05 (0.80–5.30)
2.7
34
0.85 (0.16–4.62)
CCS Subtypes
Large artery
atherosclerosis
Other uncommon
causes
Undetermined
CCS indicates Causative Classification of Stroke; CI, confidence interval; HR,
hazard ratio; and PLA, periventricular leukoaraiosis.
much as 17× more common—than symptomatic infarcts during the first few days after an ischemic stroke,24 impaired ability of the brain to tolerate ischemia and its reduced capacity to
compensate for the lost function in the presence of extensive
leukoaraiosis could potentially facilitate conversion of asymptomatic infarcts into symptomatic infarcts. In support of this
view, procedures that are associated with high risk of embolism to the brain, such as carotid artery stenting,25,26 carotid
endarterectomy,26,27 and total aortic arch replacement,28 confer
a higher perioperative risk of symptomatic stroke in patients
with extensive leukoaraiosis when compared with those with
less severe leukoaraiosis. In the International Carotid Stenting
Study (ICSS), patients with extensive leukoaraiosis (as
defined by age-related white matter changes score, ≥5) were
1.5× more likely to develop periprocedural infarction on DWI
after stenting than those with lower leukoaraiosis scores.26 In
the present study, patients with high-risk pathogeneses for
asymptomatic recurrence, such as large artery atherosclerosis
or acute arterial dissection, had 2 to 3× higher odds of symptomatic recurrence in the presence of extensive leukoaraiosis
when compared with those with lower risk pathogeneses, such
as small artery occlusion (Table 4). This finding further supports the view that the brain is more vulnerable to develop
symptomatic stroke in an event of recurrent infarction in the
presence of extensive leukoaraiosis.
The strengths of this study are large sample size, large
number of outcome events, comprehensive diagnostic assessment, rigorous identification of pathogenic stroke subtypes
using the CCS system, MRI-based assessment of leukoaraiosis, and imaging confirmation of recurrent events. There
are limitations as well. Because this was not a prospective
484 Stroke February 2014
Downloaded from http://stroke.ahajournals.org/ by guest on June 15, 2017
longitudinal study, we were not able to collect 90-day follow-up data in all of the subjects. The extent of leukoaraiosis
and independent predictors of stroke recurrence (hypertension, National Institute of Health Stroke Scale score, chronic
stroke, and CCS subtypes of large artery atherosclerosis, and
other uncommon causes), however, did not substantially differ
between patients with and without complete follow-up, arguing against selection of a particular risk population. Although
the 90-day cumulative estimate for recurrent stroke (5.9%)
was comparable with the published rates in prospective longitudinal series,29,30 we think this rate slightly underestimates
the true clinical recurrence rate in our population because we
did not consider clinical stroke events that are not accompanied by new, clinically relevant, and spatially distinct infarct
as recurrence. We note the importance of radiographic confirmation in diagnosis of clinical recurrence for future studies
that intend to replicate our results. In this study, we focused on
symptomatic events. Future studies with serial imaging could
identify whether leukoaraiosis burden is also a risk factor for
asymptomatic infarcts. We used a dichotomized visual rating
scale for leukoaraiosis. Although quantitative assessment of
leukoaraiosis load would provide more information, qualitative assessment allows the identification of high-risk individuals on the basis of readily available leukoaraiosis information
in a typical clinical setting. Although the relationship between
leukoaraiosis and early stroke recurrence was statistically significant, the correlation was not perfect. The lack of a clear
dose-dependent effect of leukoaraiosis might be because of
small number of patients in the higher leukoaraiosis load
strata. Alternatively, severe strokes occurring in the presence
of extensive leukoaraiosis would mask new symptoms in an
event of recurrent infarction and thereby lead to underestimation of symptomatic recurrence rate.1,6
Several characteristics derived from the neuroimages of
patients with acute ischemic stroke seem to predict risk of
early stroke recurrence. These include multiple acute infarcts,
simultaneous acute infarcts in both hemispheres or in both
anterior and posterior circulations, multiple infarcts of different ages (combination of acute and subacute infarcts), and
isolated cortical location of an infarct.31 We now describe a
fifth imaging predictor. The difference between leukoaraiosis
and the other imaging predictors is that the latter predictors
are primarily based on DWI. Infarcts on DWI are highly subject to changes in their number and pattern during the acute
phase. In contrast, leukoaraiosis on FLAIR images is a relatively stable and easily measurable parameter with acceptable inter- or intraobserver reliability. It requires further
studies to show whether leukoaraiosis confers prognostic
information that is not conveyed by other imaging predictors and thus provides added value in ­risk-stratification after
ischemic stroke.
Acknowledgments
We thank Dr Arsava for his help with data collection.
Sources of Funding
Dr Ay was supported by National Institutes of Health grant
R01-NS059710.
Disclosures
Dr Ay is responsible for ensuring that full disclosures appear on the
article and that the page proof reflects the disclosures listed. The other
authors report no conflicts.
References
1. Arsava EM, Rahman R, Rosand J, Lu J, Smith EE, Rost NS, et al.
Severity of leukoaraiosis correlates with clinical outcome after ischemic
stroke. Neurology. 2009;72:1403–1410.
2. Arsava EM, Bayrlee A, Vangel M, Rost NS, Rosand J, Furie KL, et al.
Severity of leukoaraiosis determines clinical phenotype after brain
infarction. Neurology. 2011;77:55–61.
3. Ay H, Benner T, Arsava EM, Furie KL, Singhal AB, Jensen MB, et al. A
computerized algorithm for etiologic classification of ischemic stroke: the
Causative Classification of Stroke System. Stroke. 2007;38:2979–2984.
4. Ay H, Arsava EM, Rosand J, Furie KL, Singhal AB, Schaefer PW, et al.
Severity of leukoaraiosis and susceptibility to infarct growth in acute
stroke. Stroke. 2008;39:1409–1413.
5. Fazekas F, Barkhof F, Wahlund LO, Pantoni L, Erkinjuntti T, Scheltens
P, et al. CT and MRI rating of white matter lesions. Cerebrovasc Dis.
2002;13(suppl 2):31–36.
6. Fazekas F, Chawluk JB, Alavi A, Hurtig HI, Zimmerman RA. MR signal
abnormalities at 1.5 T in Alzheimer’s dementia and normal aging. AJR
1987;149:351–356.
7.Fleiss JL. Measuring nominal scale agreement among many raters.
Psychol Bull 1971;76:378–382.
8. Vermeer SE, Prins ND, den Heijer T, Hofman A, Koudstaal PJ, Breteler
MM. Silent brain infarcts and the risk of dementia and cognitive decline.
N Engl J Med. 2003;348:1215–1222.
9. Fu JH, Lu CZ, Hong Z, Dong Q, Luo Y, Wong KS. Extent of white matter
lesions is related to acute subcortical infarcts and predicts further stroke
risk in patients with first ever ischaemic stroke. J Neurol Neurosurg
Psychiatry. 2005;76:793–796.
10. de Laat KF, Tuladhar AM, van Norden AG, Norris DG, Zwiers MP, de
Leeuw FE. Loss of white matter integrity is associated with gait disorders in cerebral small vessel disease. Brain. 2011;134(pt 1):73–83.
11. Liou LM, Chen CF, Guo YC, Cheng HL, Lee HL, Hsu JS, et al. Cerebral
white matter hyperintensities predict functional stroke outcome.
Cerebrovasc Dis. 2010;29:22–27.
12. Palumbo V, Boulanger JM, Hill MD, Inzitari D, Buchan AM; CASES
Investigators. Leukoaraiosis and intracerebral hemorrhage after thrombolysis in acute stroke. Neurology. 2007;68:1020–1024.
13. Wong TY, Klein R, Sharrett AR, Couper DJ, Klein BE, Liao DP, et al;
ARIC Investigators. Atheroslerosis Risk in Communities Study. Cerebral
white matter lesions, retinopathy, and incident clinical stroke. JAMA.
2002;288:67–74.
14. de Groot JC, de Leeuw FE, Oudkerk M, van Gijn J, Hofman A, Jolles J,
et al. Cerebral white matter lesions and cognitive function: the Rotterdam
Scan Study. Ann Neurol. 2000;47:145–151.
15. De Groot JC, De Leeuw FE, Oudkerk M, Van Gijn J, Hofman A, Jolles J,
et al. Periventricular cerebral white matter lesions predict rate of cognitive decline. Ann Neurol. 2002;52:335–341.
16. Silbert LC, Howieson DB, Dodge H, Kaye JA. Cognitive impairment
risk: white matter hyperintensity progression matters. Neurology.
2009;73:120–125.
17. van Straaten EC, Harvey D, Scheltens P, Barkhof F, Petersen RC, Thal
LJ, et al; Alzheimer’s Disease Cooperative Study Group. Periventricular
white matter hyperintensities increase the likelihood of progression from amnestic mild cognitive impairment to dementia. J Neurol.
2008;255:1302–1308.
18. de Leeuw FE, De Groot JC, Oudkerk M, Witteman JC, Hofman A, van
Gijn J, et al. Aortic atherosclerosis at middle age predicts cerebral white
matter lesions in the elderly. Stroke. 2000;31:425–429.
19. Smith EE, Gurol ME, Eng JA, Engel CR, Nguyen TN, Rosand J, et al.
White matter lesions, cognition, and recurrent hemorrhage in lobar intracerebral hemorrhage. Neurology. 2004;63:1606–1612.
20. Brodal P. The Central Nervous System: Structure and Function. 2nd ed.
New York, NY: Oxford University Press; 1998.
21.Gerdes VE, Kwa VI, ten Cate H, Brandjes DP, Büller HR, Stam
J; Amsterdam Vascular Medicine Group. Cerebral white matter
lesions predict both ischemic strokes and myocardial infarctions in
patients with established atherosclerotic disease. Atherosclerosis.
2006;186:166–172.
Kim et al Leukoaraiosis and Early Stroke Recurrence 485
22. Auriel E, Bornstein NM, Berenyi E, Varkonyi I, Gabor M, Majtenyi K,
et al. Clinical, radiological and pathological correlates of leukoaraiosis.
Acta Neurol Scand. 2011;123:41–47.
23. Debette S, Beiser A, DeCarli C, Au R, Himali JJ, Kelly-Hayes M, et al.
Association of MRI markers of vascular brain injury with incident stroke,
mild cognitive impairment, dementia, and mortality: the Framingham
Offspring Study. Stroke. 2010;41:600–606.
24.Kang DW, Latour LL, Chalela JA, Dambrosia J, Warach S. Early
ischemic lesion recurrence within a week after acute ischemic stroke.
Ann Neurol. 2003;54:66–74.
25. Ederle J, Brew S, Dobson J, Featherstone RL, Brown MM. Leukorariosis
and perioperative risk of stroke in patients treated for symptomatic
carotid stenosis randomised in the international carotid stenting study
(ICSS). Cerebrovasc Dis. 2009;27(suppl 6):10.
26. Gensicke H, Zumbrunn T, Jongen LM, Nederkoorn PJ, Macdonald S,
Gaines PA, et al; ICSS-MRI Substudy Investigators. Characteristics
of ischemic brain lesions after stenting or endarterectomy for
symptomatic carotid artery stenosis: results from the international
carotid stenting study-magnetic resonance imaging substudy. Stroke.
2013;44:80–86.
27. Arshad A, Altaf N, Goode S, Auer DP, MacSweeney ST. Leukoaraiosis
predicts the need for intraoperative shunt placement during carotid endarterectomy. Perspect Vasc Surg Endovasc Ther. 2009;21:173–177.
28. Morimoto N, Okada K, Uotani K, Kanda F, Okita Y. Leukoaraiosis and
hippocampal atrophy predict neurologic outcome in patients who undergo
total aortic arch replacement. Ann Thorac Surg. 2009;88:476–481.
29.Petty GW, Brown RD Jr, Whisnant JP, Sicks JD, O’Fallon WM,
Wiebers DO. Survival and recurrence after first cerebral infarction: a
­population-based study in Rochester, Minnesota, 1975 through 1989.
Neurology. 1998;50:208–216.
30. Moroney JT, Bagiella E, Paik MC, Sacco RL, Desmond DW. Risk factors for early recurrence after ischemic stroke: the role of stroke syndrome and subtype. Stroke. 1998;29:2118–2124.
31. Ay H, Gungor L, Arsava EM, Rosand J, Vangel M, Benner T, et al. A
score to predict early risk of recurrence after ischemic stroke. Neurology.
2010;74:128–135.
Downloaded from http://stroke.ahajournals.org/ by guest on June 15, 2017
Extensive Leukoaraiosis Is Associated With High Early Risk of Recurrence After Ischemic
Stroke
Gyeong-Moon Kim, Kwang-Yeol Park, Ross Avery, Johanna Helenius, Natalia Rost, Jonathan
Rosand, Bruce Rosen and Hakan Ay
Downloaded from http://stroke.ahajournals.org/ by guest on June 15, 2017
Stroke. 2014;45:479-485; originally published online December 26, 2013;
doi: 10.1161/STROKEAHA.113.003004
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2013 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/45/2/479
Data Supplement (unedited) at:
http://stroke.ahajournals.org/content/suppl/2013/12/26/STROKEAHA.113.003004.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/
Supplemental material
Supplementary Table I: Fazekas score and 90-day stroke recurrence.
Fazekas
Score
Extensive LA
Extensive PLA
Recurrence
No Recurrence
Recurrence No Recurrence
(%)
(%)
(%)
0
4 (3.8)
216 (9.5)
1
15 (14.2)
2
Extensive SLA
Recurrence
No Recurrence
(%)
(%)
(%)
9 (8.5)
313 (13.8)
14 (13.2)
426 (18.8)
308 (13.6)
44 (41.5)
1058 (46.6)
53 (50.0)
1138 (50.1)
31 (29.2)
757 (33.3)
42 (39.6)
605 (26.6)
31 (29.2)
478 (21.0)
3
18 (17.0)
352 (15.5)
11 (10.4)
296 (13.0)
8 (7.5)
230 (10.1)
4
24 (22.6)
310 (13.6)
-
-
-
-
5
11 (10.4)
165 (7.3)
-
-
-
-
6
3 (2.8)
165 (7.3)
-
-
-
-