Diagnostic Accuracy and R

isk Factors of the DifferentLacunar Syndromes

Marianne Altmann, MD,*† Bente Thommessen, MD, PhD,*

Ole Morten Rønning, MD, PhD,*† Antje S. Reichenbach, MD, PhD,*

and Brynjar Fure, MD, PhD‡x

From the *Medical Div

University Hospital, Løre

versity of Oslo, Oslo; ‡

Knowledge Centre for th

of Geriatric Medicine, Os

Received December 19

accepted March 15, 2014

Supported by the

Authority Grant number

Address corresponden

of Neurology, Medical

N-1478 Lørenskog, Norw

1052-3057/$ - see front

� 2014 by National Str

http://dx.doi.org/10.1

Journal of Stroke and C

Background: The lacunar syndrome is characterized by pure motor, pure sensory, or

sensorimotor hemisymptoms without cortical deficits. It may be less predictable for

a lacunar infarct (LI) than previously believed. The aims of the present studywere to

evaluate the diagnostic accuracy of the different lacunar syndromes and investigate

factors associated with acute LI on diffusion-weighted imaging (DWI). Methods:Consecutive patients presenting with an acute lacunar syndrome who were

admitted to the stroke unit were enrolled. The patients were examined clinically

and underwent magnetic resonance imaging. The sensitivity and specificity of the

different lacunar syndromes were assessed using DWI as reference test, and we esti-

mated positive and negative predictive values. Patients were divided into a LI group

and a group without LI. Between-group differences were analyzed by c2 test, t test,

and Mann–Whitney U test, as appropriate. Logistic regression was performed to

analyze predictors of LI. Candidate variables were pure motor syndrome, age,

gender, hypertension, precerebral or intracerebral stenosis, atrial fibrillation, dia-

betes, coronary heart disease, and smoking.Results: Eighty-six patients with lacunar

syndrome underwent DWI. The positive predictive value of the lacunar syndrome

was 65.1% and 75% for the pure motor syndrome. Of the candidate variables,

only pure motor syndrome and male gender had significant associations with LI

on imaging. Conclusions: The clinical diagnosis of patients with lacunar syndromes

is inaccurate, especially among patients with sensorimotor syndrome. DWI is

mandatory for obtaining an accurate diagnosis of the infarct. Key Words: Lacunar

syndrome—lacunar infarct—acute stroke—diffusion-weighted MRI—diagnostic

accuracy—cerebral infarction.

� 2014 by National Stroke Association

ision, Department of Neurology, Akershus

nskog; †Institute of Clinical Medicine, Uni-

Specialist Health Section, The Norwegian

e Health Services, Oslo; and xDepartment

lo University Hospital, Oslo, Norway.

, 2013; revision received February 5, 2014;

.

South-Eastern Norway Regional Health

2010091.

ce to Marianne Altmann, MD, Department

Division, Akershus University Hospital,

ay. E-mail: marianne.altmann@ahus.no.

matter

oke Association

016/j.jstrokecerebrovasdis.2014.03.014

erebrovascular Diseases, Vol. 23, No. 8 (Septem

Lacunar infarcts (LI) are small subcortical infarcts

defined as ,15 mm in the chronic phase. Typically, they

are located in the basal ganglia, thalamus, internal

capsule, corona radiata, or brainstem, and caused by oc-

clusion of a single perforating end artery deep in the

brain. LI have been related to cerebral small vessel dis-

ease, pathologically characterized by lipohyalinosis,1

which is a local process in the arterioles in the brain’s

white matter. However, recent studies2 have demon-

strated other causes of lacunar stroke, for example, cardial

and aortic embolism. Thus, the etiology of LI has been un-

der debate during the last decade.

The clinical symptoms in patients with LI found in au-

topsy studies3 are characterized by pure motor, pure

ber), 2014: pp 2085-2090 2085

M. ALTMANN ET AL.2086

sensory, or sensorimotor hemisymptoms without cortical

deficits (aphasia, apraxia, agnosia, neglect phenomena,

and so forth) or homonymous hemianopia.1 The puremo-

tor syndrome was typically associated with LIs in the in-

ternal capsule or in the brainstem, whereas the pure

sensory syndrome in most cases was explained by lesions

in the thalamus. In addition, LIs in the brainstem were

related to ataxic hemiparesis or the dysarthria–clumsy

hand syndrome. Recent studies,4,5 however, have shown

that lesions in different locations can cause the same

clinical syndrome.

The Oxfordshire Community Stroke Project (OCSP)

clinical classification was based on symptoms and signs

of different stroke syndromes, and proposed 4 defined

subgroups of cerebral infarction.6 A lacunar infarct

(LACI) was defined as an infarct confined to the territory

of the deep perforating arteries.7 However, the accuracy

of the OCSP classification8 has been poor in patients

with small infarcts, and in addition, the clinical lacunar

syndrome comprises a heterogeneous group of patients

with stroke including deep infarcts, cortical lesions, and

cerebral hemorrhages.

Stroke radiologic diagnosis has been largely improved

with early diffusion-weighted imaging (DWI), until now

the most sensitive imaging in acute ischemic stroke.9

However, using radiologic techniques as a gold standard

for LIs may be challenging, because not all acute lesions

identified as LIs on DWI cavitate and become lacunes,10,11

that is, cavities filled with cerebrospinal fluid.

In the Northern Manhattan Stroke Study Experience, a

lacunar syndrome was found to have a positive predic-

tive value (PPV) of 87% for detecting a LI on brain

imaging. However, diffusion-weighted magnetic reso-

nance imaging (DW-MRI) was not performed.12 In

another prospective study,13 patients were evaluated

with perfusion-weighted imaging/DWI, which altered

the final diagnosis from a clinical assumed lacunar infarc-

tion to a radiologic evident nonlacunar infarction in 13 of

19 patients presenting with lacunar syndromes. One

small study with 23 patients showed that almost all pa-

tients presenting with lacunar syndrome had acute le-

sions on DW-MRI, and only a minor proportion (2/23)

had cortical lesions.14 On the other hand, a study

including 111 patients with lacunar syndromes reported

that 40.5% had nonlacunar infarcts on DWI.15 One study16

showed that only 44.1% of the patients with lacunar

infarction on DWI had lacunar syndrome, and finally,

another study showed that lesions in a variety of locations

can cause the same lacunar syndrome.5 There is also an

increasing uncertainty regarding the etiology of LIs.

Hence, the value of the clinical classification of stroke

into lacunar syndromes is under debate.17-19

Inmost cases, DWI can identify an acute ischemic lesion.

However, in many stroke centers, magnetic resonance im-

aging (MRI) is not available for all patients during the

acute phase when a precise diagnosis should be made.

Consequently, knowledge about diagnostic accuracy be-

comes even more important to help the clinician make

the correct diagnosis and give the most effective treatment.

The aims of the present study were to compare clinical

lacunar syndromes with LIs on DW-MRI, find the diag-

nostic accuracy of the different lacunar syndromes in pre-

dicting LIs, and to investigate risk factors associated with

acute LIs identified on MRI.

Materials and Methods

We recruited 119 consecutive patients presenting with

an acute lacunar syndrome who were admitted to the

stroke unit of Akershus University Hospital from

February 2011 to January 2013. The patients underwent

standard examination at our stroke unit including blood

samples, electrocardiogram (ECG) records, cerebral

computed tomography (CT) at admittance, and color

duplex of precerebral and intracranial arteries. Presence

of symptomatic carotid or middle cerebral artery stenosis

greater than or equal to 50%were registered. All included

patients were examined clinically by an experienced

stroke neurologist (M.A.).

The diagnosis of lacunar syndrome was based on the

patients’ history and neurologic examination (findings

compatible with lacunar syndrome). Patients who

were treated with intravenous thrombolysis were

included, even when their symptoms lasted less than

24 hours. Exclusion criteria were intracerebral hemor-

rhage and transient ischemic attack (symptoms lasting

,24 hours). Patients who underwent only CT scan and

not DW-MRI were excluded.

Assessments

Neurologic impairment was assessed by using an 11-

item version of the National Institutes of Health Stroke

Scale (NIHSS)20 on day 1 and at discharge. Global function

was evaluated using the modified Rankin Scale (mRS)21,22

at discharge. In addition, OCSP and Barthel Activity of

Daily Living index23,24 were recorded at discharge. We

registered risk factors (hypertension, diabetes,

hypercholesterolemia, body mass index [BMI], atrial

fibrillation, coronary heart disease [previous myocardial

infarction or angina pectoris], mechanical heart valve,

smoking, and previous stroke/transient ischemic attack).

Evaluations and investigations are listed in Table 1.

Patients underwentMRIwithin aweek to identify acute

cerebrovascular lesions. The brain imaging was done on

Philips Achieva (Royal Philips, Amsterdam, The

Netherlands) 1.5T or 3T MRI scanners using standard se-

quences, using T1-weighted sagittal, T2-weighted axial,

T2/fluid attenuated inversion recovery (FLAIR) weighted

coronal, and diffusion-weighted axial imaging. Isolated

acute ischemic lesions on DWIwere defined as lacunar in-

farctions if ,15 mm and located subcortically or in the

brainstem.25

Table 1. Evaluations and investigations

Inclusion criteria Acute clinical lacunar

syndrome

Exclusion criteria Intracerebral hemorrhage

and TIA

Standard examination Blood tests

ECG

Pre-/intracerebral color duplex

Classifications NIHSS, OCSP, TOAST,

mRS, Barthel ADL index

Radiology CT at admission MRI after the

acute phase

Risk factors registration Hypertension

Diabetes

Hypercholesterolemia

Coronary heart disease

Previous atrial fibrillation

Previous stroke or TIA

Large vessel disease

Current smokers

Abbreviations: ADL, Activity of Daily Living; CT, computed to-

mography; ECG, electrocardiogram records; MRI, magnetic reso-

nance imaging; mRS, modified Rankin Scale; NIHSS, National

Institutes of Health stroke scale; OCSP, Oxfordshire Community

Stroke Project; TIA, transient ischemic attack; TOAST, Trial of

ORG 10172 in Acute Stroke Treatment.

Table 2. Sensitivity and specificity of the different lacunar

syndromes

Clinical

presentation Sensitivity Specificity PPV NPV

Motor syndrome .68 .57 .75 .49

Sensorimotor

syndrome

.23 .53 .48 .27

Motor syndrome

with triparesis

.67 .60 .77 .47

Abbreviations: NPV, negative predictive value; PPV, positive

predictive value.

DIAGNOSTIC ACCURACY OF DIFFERENT LACUNAR SYNDROMES 2087

Patients were categorized into 2 groups regardless of

their lacunar syndrome (motor, sensory, or sensorimotor):

‘‘biparesis’’ (symptoms in2 limbs or 1 limb1 facial paresis)

and ‘‘triparesis’’ (symptoms in arm and leg 1 facial

paresis).

Statistical Analyses and Ethical Aspects

The sensitivity and specificity of the different lacunar

syndromes were assessed using DWI as reference test

(‘‘gold standard’’). The values are listed in Table 2. We

also analyzed patients divided into the LI group and the

no LI group. In addition, positive and negative predictive

values were estimated, as these may be more useful in

clinical practice. Between-group differences of the risk

factors and the different syndromes were performed us-

ing 2-by-2 tables, chi-square test, and unpaired 2-sample

t test. Differences in neurologic impairment were

analyzed by Mann–Whitney U test for nonparametric

samples. Logistic regression analyses were performed to

identify factors associated with LI on DWI. Candidate

variables in the binary model were pure motor syndrome,

age, gender, hypertension, precerebral or intracerebral

stenosis, atrial fibrillation, diabetes, coronary heart dis-

ease, hypercholesterolemia, and smoking. Variables

reaching P , .20 (Table 3) were subjected to multivariate

analyses with stepwise elimination removing variables

with a significance level of P . .05. Despite a P value

..20, age and hypertension were also included in the

model, because these were considered as important vari-

ables. The results are presented as odds ratios (ORs) and

95% confidence intervals (CIs).

Data were analyzed using SPSS version 19 (SPSS Inc,

Chicago, IL). Oral and written informed consent was ob-

tained. The study was approved by The Regional Com-

mittee for Ethics in Medical Research.

Results

A total of 119 patients presenting with a lacunar syn-

drome were recruited. Because of capacity problems in

MRI scanning, 33 patients underwent only CT scanning

and were excluded, see Figure 1. The baseline characteris-

tics and risk factors of these patients did not differ from

that of the included patients, except for a significantly

higher prevalence of diabetes and higher proportion of

women. Eighty-six patients underwent both CT and

MRI scanning. Of these, 51 patients had pure motor syn-

drome, 5 had pure sensory syndrome, and 27 had senso-

rimotor syndrome. One patient had an ataxic hemiparesis

and 2 patients had dysarthria–clumsy hand syndrome,

see Figure 2.

The mean age was 69.3 years (standard

deviation 5 12.1), and 65.1% of the patients were men.

The median NIHSS score was 3 (IQR 2-4) at admission

and 1.5 (IQR 0-3) at discharge, whereasmedianmRS score

was 2 (IQR 1-3) at discharge and median Barthel Activity

of Daily Living index (day 2-4) was 20 (IQR 16-20). Nine

patients (10.5%) were treated with intravenous thrombol-

ysis. Characteristics and vascular risk factors are pre-

sented in Table 4.

Sixty-nine patients (80.2%) had an acute ischemic lesion

on DWI, only 6 (8.7%) of these appeared on the initial CT.

Fifty-six patients (65.1%) had a lacunar lesion, 13 patients

(15.1%) had 1 or more nonlacunar lesions. Seventeen pa-

tients had no sign of acute infarct. If we use DWI as a

gold standard for having lacunar infarction, the PPV of

the lacunar syndrome is 65.1% (56/86). The sensitivity,

specificity, positive and negative predictive value of the

pure motor syndrome and sensorimotor syndrome are

listed in Table 2. We did no analyses for the pure sensory

syndrome, ataxic hemiparesis, or dysarthria–clumsy

Table 3. Logistic regression of predictors of lacunar infarct on DWI

Bivariate analyses Multivariate analyses

OR (95% CI) P value

Adjusted OR

(95% CI) P value

Pure motor syndrome 2.76 (1.11-6.89) .029 2.66 (1.04-6.80) .041

Age 1.02 (.98-1.06) .27 - -

Male 2.73 (1.08-6.91) .034 2.63 (1.01-6.83) .047

Hypertension .70 (.29-1.70) .43 - -

Smoking 1.85 (.74-4.59) .19 - -

Coronary disease 1.67 (.54-5.18) .38 - -

Atrial fibrillation .92 (.25-3.36) .90 - -

Large vessel disease 1.09 (.23-3.53) .89 - -

Hypercholesterolemia .91 (.32-2.58) .86 - -

Diabetes .71 (.20-2.48) .59 - -

Abbreviations: CI, confidence interval; DWI, diffusion-weighted imaging; OR, odds ratio.

M. ALTMANN ET AL.2088

hand syndrome, because very few patients presented

with those syndromes. Forty-one patients had ‘‘biparesis’’

and 45 had ‘‘triparesis’’ (symptoms in arm and leg1 facial

paresis). We calculated the sensitivity and PPV for

patients presenting with motor syndrome and triparesis,

but this did not change the values. In Table 3, risk factors

in patients with lacunar, nonlacunar, or no lesions on MRI

and values for triparesis are listed.

The Pearson chi-square test, t test, and Mann–Whitney

U test were used when appropriate, to compare risk fac-

tors and characteristics between the 2 groups, to look for

significant differences. The results are presented in

Table 4. There were no significant differences between

the groups regarding vascular risk factors. The proportion

of male patients and pure motor syndrome were signifi-

cantly higher in the LI group (P 5 .031 and P 5 .027).

In thebi- andmultivariate logistic regressionmodel, pure

motor syndrome and male gender were the only variables

with significant associations with LI on DWI, see Table 3.

Discussion

In the present study, a lacunar syndrome has an overall

low PPV (65.1%) for predicting an acute LI on DWI. This

Patients presenting withlacunar syndrome

N=119

Only CT

N=33

CT and DW-MRI

N=86

Negative DWI

N=17

LacunarinfarctN=56

No acuteischemic infarct

N=14

Acuteischemic infarct

N=19

Non-lacunarinfarctN=3

LacunarinfarctN=16

Acuteischemic infarct

N=69

Non-lacunarinfarctN=13

Figure 1. Enrollment diagram.

is in accordance with previous DW-MRI studies.8,9 The

PPV was particularly low among patients with a

sensorimotor syndrome. Previous reports,12,26,27 which

showed a high diagnostic accuracy of the lacunar

syndrome, did not use DWI, but only CT or

conventional MRI. In these studies, patients with no

verified infarct on CT/MRI were either excluded or

classified as lacunar infarction,12,26,28 which may have

resulted in a falsely elevated PPV of the lacunar

syndrome.

Recent studies using DWI have demonstrated that the

OCSP classification has a particularly low PPV in predict-

ing infarct location of small infarcts.8,15 One of these

studies found a PPVof LACI as low as 39%.8 These results

demonstrate that the OCSP classification does not permit

accurate discrimination between lacunar and small

cortical infarcts. This is in accordance with the findings

in our study. In the Bergen Stroke Study,15 40.5% of the pa-

tients presenting with lacunar syndrome had nonlacunar

infarcts on DWI. This study did not include patients who

had no infarct on DWI. In our study, 17 patients had no

recent ischemic lesion on DWI at all. Four of these had

Figure 2. Clinical presentation and identification of an acute cerebrovas-

cular lesion on diffusion-weighted imaging.

Table 4. Characteristics and risk factors

LI,

N 5 56

NLI,

N 5 30

Age, mean (SD) 70.4 (11.4) 67.33 (13.4)

Males 41 (73.2) 15 (50.0)

Current smokers 18 (32.0) 14 (46.7)

Hypertension 33 (58.9) 15 (50.0)

Diabetes 10 (17.9) 4 (13.3)

Hypercholesterolemia 42 (75.0) 23 (76.7)

Coronary disease* 14 (25.0) 5 (16.7)

Large vessel diseasey 10 (17.9) 5 (16.7)

Previous atrial fibrillation 8 (14.3) 4 (13.3)

Atrial fibrillation at

admission

5 (8.9) 4 (13.3)

Previous stroke or TIA 10 (17.9) 7 (23.3)

NIHSS at admission,

median (IQR)

3 (2-4) 3 (1-4)

NIHSS at discharge,

median (IQR)

2 (.25-3) 1 (0-2)

Barthel ADL index

day 2-4, median (IQR)

20 (15-20) 20 (17-20)

mRS at discharge,

median (IQR)

2 (1-3) 1.5 (.75-2.0)

Treated with intravenous

thrombolysis

2 (3.6) 7 (23.3)

Abbreviations: ADL, Activities of Daily Living; IQR, interquar-

tile range; LI, lacunar infarct (group); mRS, modified Rankin Scale;

NIHSS, National Institutes of Health Stroke Scale; NLI, no lacunar

infarct (group); SD, standard deviation; TIA, transient ischemic

attack.

Results are given as n (%) unless indicated otherwise.

*Previous myocardial infarction and/or angina pectoris.

y.50% stenosis in the internal carotid artery or middle cerebral

artery.

DIAGNOSTIC ACCURACY OF DIFFERENT LACUNAR SYNDROMES 2089

been given intravenous thrombolytic treatment, which

may explain the absence of an ischemic lesion. The re-

maining 13 patients constitute 15.1% of all patients, which

is more than in earlier reports.8,29 Although DWI has a

high sensitivity for detecting acute ischemic lesions,30

false-negative DWI cases do occur. There might be

ischemic lesions that are not visible on DWI, and this is

particularly the case for small ischemic lesions, especially

LIs and infarcts located in the brain stem.29 In our sample

with lacunar syndromes, most of the patients had minor

stroke, and we would expect a higher proportion of

false-negative DWIs. Some of the 13 patients might have

had a nonischemic diagnosis (a stroke mimic) such as

migraine, functional paresis, and so forth.31 In clinical

practice, these differential diagnoses may be challenging.

Given the low PPV of the lacunar syndrome, is the

lacunar syndrome useless in providing the correct diag-

nosis in everyday clinical practice? The pure motor syn-

drome was the most common of the lacunar syndromes

(59.3%) in the present study and had the highest PPV

(75%). Still 1 of 4 patients with pure motor syndrome

had not suffered a LI. The sensorimotor syndrome had a

PPV of 48%, which is no better than chance. MRI is not

available for all patients with stroke. The probable extent

and localization of the ischemic lesion then must be based

on clinical findings. Determining the etiology requires in-

vestigations to search for cardial and aortic embolism,

that is, 24-hour ECG and transesophageal Doppler. The

use of DW-MRI would have added information regarding

single or multiple lesions. Multiple new lesions strongly

indicate large vessel or embolic etiology.25,32

According to Fisher’s definition of the pure motor

syndrome, the patients should present with a paresis

involving the face, arm, and leg (‘‘triparesis’’). Whenwe re-

classified our patients using this strict definition of the

pure motor syndrome, the values of sensitivity, specificity,

and PPV did not improve. The incidence of a pure motor

syndrome was significantly higher in the LI group than

in the group without LI. The pure motor syndrome has

the highest diagnostic accuracy among the lacunar syn-

dromes analyzed. No significant differences in risk factors

were found between the groups, which are in accordance

with other studies.16,26 Hypertension, smoking, and

diabetes are important but rather nonspecific risk factors

for LI, and do not differ from other stroke subtypes. If

we compare the 2 groups of patients with an acute

ischemic lesion on DWI, lacunar or nonlacunar, the latter

are older and have a higher frequency of atrial

fibrillation at admission and large vessel stenosis (not

statistically significant). We only included patients with

lacunar syndrome, and therefore, the number of patients

with cortical lesions was probably too low to reach

significant differences. Others have reported significantly

higher proportion of large vessel disease or atrial

fibrillation among patients with cortical infarcts.15,33

In the present study, we classified only 1 patient with a

classical ataxic hemiparesis. It is difficult to distinguish

between the ataxic hemiparesis and the pure motor syn-

drome. Even a small paresis may lead to ataxic move-

ments in the leg and arm. Asdaghi et al8 found that 70%

of the patients diagnosed with an ataxia–hemiparesis

lacunar syndrome had cortical MRI lesions. According

to the NIHSS, ataxia should represent cerebellar ataxia

without paresis, which probably is a more feasible defini-

tion in clinical practice.

The prospective design of our study including consecu-

tive patients with lacunar syndrome reflects the real life

experience in a stroke unit. The study has demonstrated a

low diagnostic accuracy of the clinical lacunar syndrome.

It is important to be aware of this uncertainty in clinical

practice. The clinical diagnosis of patientswith lacunar syn-

dromes is inaccurate, especially among patients with non-

motor syndromes. DWI is mandatory for obtaining an

accurate diagnosis of the infarct. Without DWI, patients

with stroke shouldundergo the same investigations regard-

lessof their clinical presentation. Further studies areneeded

to clarifywhether lacunar strokes shownonDWI can justify

a limited search of etiology.

M. ALTMANN ET AL.2090

References

1. Fisher CM. Lacunar strokes and infarcts: a review.Neurology 1982;32:871-876.

2. Caso V, Budak K, Georgiadis D, et al. Clinical significanceof detection of multiple acute brain infarcts on diffusionweighted magnetic resonance imaging. J Neurol Neuro-surg Psychiatr 2005;76:514-518.

3. Fisher CM. Lacunes: small, deep cerebral infarcts.Neurology 1965;15:774-784.

4. Hiraga A, Uzawa A, Kamitsukasa I. Diffusion weightedimaging in ataxic hemiparesis. J Neurol Neurosurg Psy-chiatr 2007;78:1260-1262.

5. Schonewille WJ, Tuhrim S, Singer MB, et al. Diffusion-weighted MRI in acute lacunar syndromes. Aclinical-radiological correlation study. Stroke 1999;30:2066-2069.

6. Bamford J, Sandercock P, Dennis M, et al. Classificationand natural history of clinically identifiable subtypes ofcerebral infarction. Lancet 1991;337:1521-1526.

7. Lindley RI, Warlow CP, Wardlaw JM, et al. Interobserverreliability of a clinical classification of acute cerebralinfarction. Stroke 1993;24:1801-1804.

8. Asdaghi N, Jeerakathil T, Hameed B, et al. Oxfordshirecommunity stroke project classification poorly differenti-ates small cortical and subcortical infarcts. Stroke 2011;42:2143-2148.

9. Lee LJ, Kidwell CS, Alger J, et al. Impact on stroke sub-type diagnosis of early diffusion-weighted magneticresonance imaging and magnetic resonance angiography.Stroke 2000;31:1081-1089.

10. Moreau F, Patel S, LauzonML, et al. Cavitation after acutesymptomatic lacunar stroke depends on time, location,and MRI sequence. Stroke 2012;43:1837-1842.

11. Potter GM, Doubal FN, Jackson CA, et al. Counting cavi-tating lacunes underestimates the burden of lacunarinfarction. Stroke 2010;41:267-272.

12. Gan R, Sacco RL, Kargman DE, et al. Testing the validityof the lacunar hypothesis: the Northern ManhattanStroke Study experience. Neurology 1997;48:1204-1211.

13. Gerraty RP, Parsons MW, Barber PA, et al. Examining thelacunar hypothesis with diffusion and perfusion mag-netic resonance imaging. Stroke 2002;33:2019-2024.

14. Lindgren A, Staaf G, Geijer B, et al. Clinical lacunar syn-dromes as predictors of lacunar infarcts. A comparison ofacute clinical lacunar syndromes and findings ondiffusion-weighted MRI. Acta Neurol Scand 2000;101:128-134.

15. Naess H, Brogger JC Jr, Idicula T, et al. Clinical presenta-tion and diffusion weighted MRI of acute cerebral infarc-tion. The Bergen Stroke Study. BMC Neurol 2009;9:44.

16. Seifert T, Enzinger C, Storch MK, et al. Acute smallsubcortical infarctions on diffusion weighted MRI:

Clinical presentation and aetiology. J Neurol NeurosurgPsychiatr 2005;76:1520-1524.

17. Jackson C, Sudlow C. Are lacunar strokes reallydifferent? A systematic review of differences in risk factorprofiles between lacunar and nonlacunar infarcts. Stroke2005;36:891-901.

18. Landau WM, Nassief A. Editorial comment–time to burnthe TOAST. Stroke 2005;36:902-904.

19. Toni D, Sacchetti ML, PrencipeM. Early clinical diagnosisof lacunar strokes. Stroke 2008;39:e152. author reply e153.

20. Goldstein LB, Samsa GP. Reliability of the national insti-tutes of health stroke scale. Extension to non-neurologistsin the context of a clinical trial. Stroke 1997;28:307-310.

21. Rankin J. Cerebral vascular accidents in patients over theage of 60. II. Prognosis. Scott Med J 1957;2:200-215.

22. van Swieten JC, Koudstaal PJ, Visser MC, et al. Interob-server agreement for the assessment of handicap instroke patients. Stroke 1988;19:604-607.

23. Mahoney FI, Barthel DW. Functional evaluation: the bar-thel index. Md State Med J 1965;14:61-65.

24. Collin C, Wade DT, Davies S, et al. The barthel ADL in-dex: a reliability study. Int Disabil Stud 1988;10:61-63.

25. Wessels T, Wessels C, Ellsiepen A, et al. Contribution ofdiffusion-weighted imaging in determination of strokeetiology. AJNR Am J Neuroradiol 2006;27:35-39.

26. Boiten J, Lodder J. Lacunar infarcts. Pathogenesis andvalidity of the clinical syndromes. Stroke 1991;22:1374-1378.

27. Melo TP, Bogousslavsky J, van Melle G, et al. Pure motorstroke: a reappraisal. Neurology 1992;42:789-795.

28. Phillips SJ, Dai D, Mitnitski A, et al. Clinical diagnosis oflacunar stroke in the first 6 hours after symptom onset:analysis of data from the glycine antagonist in neuropro-tection (GAIN) Americas trial. Stroke 2007;38:2706-2711.

29. Sylaja PN, Coutts SB, Krol A, et al. When to expect nega-tive diffusion-weighted images in stroke and transientischemic attack. Stroke 2008;39:1898-1900.

30. Schellinger PD, Bryan RN, Caplan LR, et al. Evidence-based guideline: the role of diffusion and perfusionMRI for the diagnosis of acute ischemic stroke: reportof the therapeutics and technology assessment subcom-mittee of the American Academy of Neurology.Neurology 2010;75:177-185.

31. Fernandes PM, Whiteley WN, Hart SR, et al. Strokes:mimics and chameleons. Pract Neurol 2013;13:21-28.

32. Wessels T, Rottger C, Jauss M, et al. Identification ofembolic stroke patterns by diffusion-weighted MRI inclinically defined lacunar stroke syndromes. Stroke2005;36:757-761.

33. Mead GE, Lewis SC, Wardlaw JM, et al. Shouldcomputed tomography appearance of lacunar stroke in-fluence patient management? J Neurol Neurosurg Psy-chiatr 1999;67:682-684.