Content uploaded by Timothy C Y Kwok
Author content
All content in this area was uploaded by Timothy C Y Kwok
Content may be subject to copyright.
PAPER
Cognitive impairment and functional outcome after stroke
associated with small vessel disease
V C T Mok, A Wong, W W M Lam, Y H Fan, W K Tang, T Kwok, A C F Hui, K S Wong
...............................................................................................................................
See end of article for
authors’ affiliations
.......................
Correspondence to:
Dr V C T Mok, Department
of Medicine, Prince of
Wales Hospital, Shatin,
The Chinese University of
Hong Kong, Hong Kong,
China; vctmok@
netvigator.com
Received 25 March
In revised form
20 July 2003
Accepted
7 September 2003
.......................
J Neurol Neurosurg Psychiatry 2004;75:560–566. doi: 10.1136/jnnp.2003.015107
Objectives: Although stroke associated with small vessel disease (SSVD) can induce both motor and
cognitive impairment, the latter has received less attention. We aimed to evaluate the frequency of the
varying severity levels of cognitive impairment, the determinants of severe cognitive impairment, and the
association of cognitive impairment with functional outcome after SSVD.
Methods: Consecutive patients admitted to hospital because of SSVD were assessed at 3 months after
stroke. We performed a semi-structured clinical interview to screen for cognitive symptoms. Severity of
cognitive symptoms was graded according to the Clinical Dementia Rating Scale (CDR). Performance on
psychometric tests (Mini-Mental State Examination, Alzheimer’s Disease Assessment Scale (cognition
subscale), Mattis Dementia Rating Scale (initiation/perseverence subscale; MDRS I/P)) of patients of
different CDR gradings was compared with that of 42 healthy controls. Basic demographic data, vascular
risk factors, stroke severity (National Institute of Health Stroke Scale; NIHSS), pre-stroke cognitive decline
(Informant Questionnaire on Cognitive Decline in the Elderly; IQCODE), functional outcome (Barthel index;
BI), Instrumental Activities Of Daily Living; IADL), and neuroimaging features (site of recent small infarcts,
number of silent small infarcts, white matter changes) were also compared among the groups. Regression
analyses were performed to find predictors of severe cognitive impairment and poor functional outcome.
Results: Among the 75 included patients, 39 (52%) complained of cognitive symptoms. The number of
patients in each CDR grading was as follows: 39 (52%) had a CDR of 0, 26 (34.7%) had a CDR of 0.5, 10
(13.3%) had a CDR of >1. Pre-stroke IQCODE and previous stroke predicted CDR>1. The NIHSS was
associated with more impaired BI. The NIHSS and MDRS I/P contributed most to impaired IADL.
Conclusions: Half of the patients with SSVD complained of varying severity of cognitive problems
3 months after stroke. Pre-stroke cognitive decline and previous stroke predict severe cognitive impairment
post stroke. Stroke severity and executive dysfunction contribute most to a poor functional outcome.
Small, deep infarcts, also known as lacunar infarcts,
and subcortical white matter changes (WMC) are
manifestations of underlying cerebral small vessel
disease.
12
Apart from sensorimotor disturbance, these
ischaemic lesions commonly induce cognitive impairment
due to disruption of the frontal–subcortical or medial
temporal limbic circuits.
3–5
The pattern of cognitive impair-
ment commonly involves executive dysfunction, memory
loss, or aphasia.
167
The progression of the cognitive
impairment can be insidious, stepwise, or both.
8–10
Eventually, cognitive impairment in small vessel disease
may progress from mild to severe cognitive impairment.
8
Subcortical vascular dementia is the current terminology
for severe cognitive impairment associated with small
vessel disease.
1
Given its strong vascular component, it is
believed to be more preventable than dementia of the
Alzheimer’s type.
11
Stroke is a common presentation of small vessel disease.
Most studies on stroke associated with small vessel
disease (SSVD) focus on motor impairment and mortal-
ity.
12–15
Few studies have investigated the frequency of
differering severity levels of cognitive impairment.
Furthermore, the influence of cognitive impairment on
the functional outcome after SSVD has rarely been
studied. The objectives of this study were to evaluate:
(a) the frequency and features of the varying severity
levels of cognitive impairment, (b) the determinants for
severe cognitive impairment, and (c) the association
between cognitive impairment and functional outcome after
SSVD.
METHODS
We evaluated consecutive patients with or without previous
stroke who were admitted to the acute stroke unit of the
Prince of Wales Hospital because of stroke or transient
ischaemic attack (TIA) from January to June 2002 (n = 294).
Brain computed tomography (CT) was performed on all
patients within 24 hours of admission. Patients with intra-
cerebral haemorrhage were excluded (n = 37). T1 and T2
weighted magnetic resonance imaging (MRI) (1.5T scanner,
Sonata; Siemens Medical System, Erlangen, Germany) and
magnetic resonance angiography (MRA) were performed on
234 patients within 3 months of their stroke. A small infarct
was defined as a well circumscribed lesion giving a
hyperintense (T2 weighted) or hypointense (T1 weighted)
signal on MRI, or a hypodense signal on CT, 0.2–2 cm size in
all dimensions, located in the subcortical white and gray
matter, cerebellar white matter, and brain stem. Small
infarcts that could account for the present neurological
deficits were considered to be relevant. WMC was defined on
MRI as ill defined hyperintense areas >5 mm on T2 weighted
................................................................
Abbreviations: BI, Barthel Index; CDR, Clinical Dementia Rating Scale;
CT, computed tomography; IADL, Instrumental Activities Of Daily Living;
IQCODE, Informant Questionnaire on Cognitive Decline in the Elderly;
MDRS I/P, Mattis Dementia Rating Scale (initiation/perseverence
subscale); MRA, magnetic resonance angiography; MRI, magnetic
resonance imaging; NIHSS, National Institute of Health Stroke Scale;
SSVD, stroke associated with small vessel disease; TCD, transcranial
Doppler; TIA, transient ischaemic attack; WMC, white matter changes
560
www.jnnp.com
imaging, and on CT as ill defined and moderately hypodense
areas of >5 mm. Transcranial Doppler (TCD) ultrasound
(EME TC-2000) was performed to screen for intracranial
large artery stenosis among patients with CT showing
relevant small infarct but upon whom MRA could not be
performed. Carotid duplex ultrasound (Philip SD800 ultra-
sound machine, 7.5 MHz transducer) was also performed
upon all patients with relevant small infarct to screen for
extracranial carotid artery stenosis.
Patients with SSVD were potentially eligible for the study.
We defined SSVD as patients having relevant small infarct
with or without WMC. We excluded patients with relevant
small infarcts that were associated with relevant intracranial
large artery disease, extracranial carotid artery stenosis,
cardiac embolic sources, and other miscellaneous causes.
Intracranial large artery disease was considered to be relevant
if the stenosis was at least moderate (>50%) in severity and
if the infarct was located within its arterial supply. The
methods of defining moderate stenosis via MRA and TCD
have been described previously.
16 17
Extracranial carotid
artery stenosis was considered to be relevant if the stenosis
was >50% and the infarct was located within its arterial
distribution. Patients were presumed to have cardiac embolic
sources if they had concurrent presence or past history of
atrial fibrillation (AF), sick sinus syndrome, metallic heart
valves, acute congestive failure, recent ((6 weeks before
stroke) myocardial infarction, atrial myxoma, or patent
foramen ovale. Electrocardiography was performed on all
patients. Transthoracic or transoesophageal echocardiogra-
phy was not required to confirm the source of emboli.
Patients were considered to have miscellaneous causes if they
had the following diseases: inflammatory disorders (for
example, systemic lupus erythematosus), carotid or vertebral
artery dissection as suggested by history and vascular
neuroimaging, recreational drug misuse (for example,
cocaine), or haematological disorders (for example, throm-
bocytosis). We particularly screened for these conditions
during the follow up at 3 months by the same neurologist
(VCTM) via clinical history, laboratory investigations, and
medical case notes.
Other exclusion criteria were: (a) cortical or large
subcortical (.2 cm) infarcts; (b) intracerebral haemorrhage;
(c) clinical signs that could not be explained by the small
infarct; (d) normal imaging; (e) non-ischaemic lesions, for
example, tumour or demyelination; (f) presence of relevant
small infarcts but with unknown vascular aetiology because
of absence of MRA, transcranial sonographic temporal
window, or default imaging appointment; (g) known pre-
existing dementing illnesses that were not due to Alzheimer’s
disease or vascular dementia, for example, chronic alcohol-
ism; (h) major depression according to the Diagnostic and
Statistical Manual of Mental Disorders, 4th edition;
18
and (i)
communication problems hindering participation in cognitive
assessment, such as a language barrier, or severe visual or
hearing loss. Patients with post-stroke aphasia were not
excluded from our study. A qualified psychiatrist (W K Tang)
excluded major depression 3 months after stroke.
Among all included patients, sites of the relevant small
infarct were classified into cerebral white matter (corona
radiata and centrum semi ovale), striatocapsule, thalamus,
cerebellum, and brainstem, based on neuroimaging
findings with reference to the clinical presentation. The
total number of silent small infarcts and severity of WMC
were also recorded. Severity of WMC was graded according
to the method reported by Wahlund et al.
19
Briefly, WMC
was graded in five different brain regions, frontal, parieto-
occipital, temporal, basal ganglia, and infratentorial,
based on T2 weighted MRI or CT. The score for each
region varied from 0 (no lesions) to 3 (diffuse involvement).
The total score was the sum of the grade given for each
region. The same neurologist (YHF) graded the WMC,
and another radiologist (WWML) provided other details of
the neuroimaging while blinded to the patients’ cognitive
status.
Cognitive assessment
The same neurologist (VCTM) performed a semi-structured
clinical interview on all patients to assess post-stroke
cognition at 3 months after the stroke. During the semi-
structured clinical interview, both the patients and close
informants were asked separately whether the patients noted
any cognitive symptoms after the index stroke. Following this
open ended question, specific questions were asked to screen
for the presence of the following cognitive symptoms:
memory loss, slow thinking (delay in decision making or
answering questions), disorientation (in time, place, or
person), aphasia (word finding difficulties, lack of fluency
in speech, or difficulties in comprehension), and disinhibi-
tion (impulsive or inappropriate behaviour). Any types of
cognitive symptoms that still remained in the week before
the interview were considered to be relevant. Cognitive
symptoms that increased following the stroke were also
noted. The interference of the cognitive symptoms on
patients’ various aspects of daily function was then graded
according to the CDR
20 21
during the same interview. The CDR
evaluated patient’s performance in six categories of cognitive
functioning: memory, orientation, judgement and problem
solving, community affairs, home and hobbies, and personal
care. The CDR scoring system was based on Morris.
21
There
are five grades in the CDR: 0, 0.5, 1, 2, and 3; the higher the
grade, the greater the cognitive impairment. In grading
impairment in each of the categories, care was taken to grade
only those impairments that were attributed to cognitive
symptoms, not to motor or mood disturbances. Apart from
the semi-structured clinical interview, the following psycho-
metric tests were also administered to all patients: Chinese
versions of the Mini-Mental State Examination (MMSE),
22
Alzheimer’s Disease Assessment Scale (cognition; ADAS-
cog),
23
and the Mattis Dementia Rating Scale (initiation/
perseverance subscale; MDRS I/P).
24
The MMSE served as a
brief global cognitive measure. The ADAS-cog is a more
extensive measure than MMSE. It evaluates memory,
orientation, language, construction, and ideational apraxia.
The MDRS I/P has been used in other studies as a brief
evaluation of executive dysfunction.
525
The same research
assistant (AW) administered the psychometric tests on all
patients while blinded to the patients’ CDR grading. The
control group for the performance of psychometric tests
comprised 42 healthy elderly people who were recruited from
the following sources: community elderly centre (n = 31),
relatives of hospital staff (n = 6), and spouses of patients
(n = 3), and hospital staff (n = 2). The criteria for controls
were absence of neurological or psychiatric diseases, or any
cognitive symptoms, independence in daily functioning, and
a CDR of 0. Again, the Chinese versions of the MMSE, ADAS-
cog, MDRS I/P, BI, and IADL were administered. All controls
and patients gave their informed consent to participate in this
study.
Pre-stroke cognitive decline was assessed via the Chinese
version of the Informant Questionnaire on Cognitive Decline
in the Elderly (IQCODE)
26
to the patients’ closest available
relatives within 2 weeks after the onset of the index stroke.
The IQCODE contains 26 items that rate changes in memory
and other cognitive abilities that have occurred in the
previous 10 years. The informants were asked by a same
research assistant to rate the changes on a 5 point scale
ranging from ‘‘much better’’ to ‘‘much worse’’.
Cognitive impairment and functional outcome after SSVD 561
www.jnnp.com
Clinical profile and functional outcome
Data on age, gender, years of education, and vascular risk
factors including hypertension (HT), diabetes mellitus (DM),
heart diseases (coronary artery disease, heart failure, or atrial
fibrillation), previous stroke or TIA, smoking, alcohol intake,
and fasting lipids level were collected during the acute
admission. Hypercholesterolaemia was defined as total
cholesterol of .6.2 mmol/L or low-density lipoprotein of
.4.1 mmol/L. We performed the National Institute of Health
Stroke Scale (NIHSS) during the first few days after stroke to
measure patients’ worst stroke severity. It mainly evaluates
motor impairment (weakness, ataxia), sensory loss, visual
field defects, and cortical deficits (aphasia, neglect). We
determined the functional outcome of the patients at
3 months after stroke using the Barthel (BI) Index and
Lawton IADL.
27
Two research nurses who were blinded to
patients’ CDR grading and NIHSS score administered the
scales. The BI and IADL evaluate basic and complex ADL
respectively. The total score for BI is 20, with a higher score
representing better function. The score for each item of the
IADL ranges from 0 to 3, with a lower score representing
better function. As some items in the IADL were not scored
because these items were not applicable to all patients owing
to personal habits or motor impairment, we took the average
of all scored items as the final IADL score.
Statistical analysis
Student’s ttest and x
2
test were used for comparing
continuous and dichotomous variables respectively between
patients and controls. Analysis of covariance, using age and
education as covariates, was used to compare psychometric
scores of controls and patients. To assess the frequency of the
varying severity of cognitive impairment, we divided the
patients into three groups based on CDR: 0, 0.5, and >1. We
grouped those with CDRs of 1, 2, and 3 into one group for
statistical analysis as there were only few patients who had
CDR >1. Analysis of variance was used to compare all
continuous variables of controls and the three cognitive
groups. Post hoc tests with Bonferroni adjustment were
performed with each analysis of variance. Because the neuro-
imaging scores were not normally distributed, Mann-
Whitney U tests with Bonferroni correction were used for
comparisons between the three cognitive groups. Variables
that were significantly different (p,0.05) between patients
with CDR >1 and (0.5 in the binomial univariate analysis
were entered into a forward stepwise multivariate binomial
regression analysis to find independent predictors of CDR >1.
We also performed two univariate linear regression analyses
to find the variables that accounted for the variance in BI and
IADL respectively. Variables that were identified as signifi-
cant in the univariate analysis (p,0.05) were entered into
the forward stepwise multivariate linear regression analyses
to examine their independent contributions to the variance of
both BI and IADL.
RESULTS
Among the 294 admitted patients, 106 had relevant small
infarcts. Among these 106 patients, 20 were excluded because
of underlying relevant intracranial large artery disease
(n = 14) and unknown vascular aetiology (n = 6). Two of
the 14 patients with intracranial large artery disease also had
concurrent relevant extracranial carotid artery disease. No
relevant small infarcts were related to cardioembolism, lone
relevant extracranial carotid disease, or other miscellaneous
causes in our present cohort. Among the 86 patients with
SSVD, 11 more patients were excluded from further cognitive
assessment because of severe depression (n = 3), death
(n = 1), default follow up (n = 4), chronic alcoholism
(n = 1), and language barrier (n = 2). Seventy five patients
with SSVD (25.5%) underwent cognitive assessment and
were included in the final analysis. Comparison of the 75
included patients with the 17 excluded patients with relevant
small infarcts but who had unknown vascular aetiology
(n = 6) or who did not undergo cognitive assessment
(n = 11) showed that the excluded patients were older
(77 years) and had more severe stroke (NIHSS = 9.1) than
those who were included in the study (70.7 years old,
p = 0.006; NIHSS = 4.3, p = 0.025). Comparisons of controls
and included patients showed no difference in terms of basic
demographic data. The performance on MMSE, ADAS-cog,
and MDRS I/P of the patients was significantly worse than
that of controls after adjustment for age and education
(table 1).
Among the 75 patients, 39 (52%) complained of cognitive
symptoms post-stroke. These cognitive symptoms repre-
sented either new onset or further increase in pre-stroke
cognitive symptoms among 32 (82.1%) of these 39 patients.
The majority of the patients (n = 36; 92.3%) complained of
memory problems. Slow thinking, disorientation, and apha-
sia occurred in 29 (74.4%), 7 (17.9%), and 7 (17.9%) patients
respectively. Disinhibition was noted only in one patient
(2.6%). No patients complained of other cognitive symptoms,
such as neglect or apraxia, in response to the open ended
question. Twenty nine patients (74.4%) had more than one
type of cognitive symptoms. Seven patients (17.9%) had
isolated memory complaints. Three patients (7.7%) com-
plained of slow thinking without associated memory
complaints or other cognitive symptoms. No patients had
isolated aphasia or disorientation. Among patients with
SSVD, the number of patients in each CDR grading was as
follows: 39 (52%) had CDR of 0, 26 (34.7%) had CDR of 0.5, 8
(10.7%) had CDR of 1, 1 (1.3%) had CDR of 2, and 1 (1.3%)
had CDR of 3. The three patients who complained of isolated
slow thinking were classified as CDR of 0 according to the
rating algorithm.
21
Comparison of patients with varying CDR grading revealed
that patients with CDR>1 had significantly more impaired
IADL, pre-stroke IQCODE, and performance on psychometric
tests than those with CDR,1, even though the groups had
similar basic demographic data, clinical features, sites of
recent infarcts, number of silent infarcts, and BI (table 2).
Although patients with CDR>1 also had more WMC and
previous stroke than patients with CDR ,1, these differences
did not reach statistical significance. Patients with CDR of 0.5
had significantly more impaired pre-stroke IQCODE than
those with CDR of 0. Patients with CDR of 0 performed
significantly worse on MDRS I/P and ADAS-cog than did
controls even though they had similar basic demographic
features.
Univariate regression analysis showed that pre-stroke
IQCODE, NIHSS, previous stroke, and WMC were associated
Table 1 Comparison of controls and patients
with SSVD
Controls
(n = 42)
Patients
(n = 75) p
Gender (% F) 58.5 48 0.549
Age 69.6 (9.9) 71.0 (11.2) 0.095
Education 5.4 (4.6) 4.8 (4.1) 0.086
BI 20 (0) 18.3 (3.5) ,0.001
IADL 0.1 (0.2) 0.9 (0.8) ,0.001
MMSE 27.7 (2.1) 24.8 (4.5) ,0.001
MDRS I/P 34.1 (3.0) 28.5 (6.4) ,0.001
ADAS-cog 9.5 (3.4) 16.5 (9.3) ,0.001
x
2
= 0.49.
F, female.
Figures are mean (SD) except for gender.
562 Mok,Wong,Lam,etal
www.jnnp.com
with CDR>1(table 3). Multivariate regression analysis only
showed that pre-stroke IQCODE (odds ratio (OR) = 95.52,
95% confidence interval (CI) =7.37 to 1248.42, p,0.001) and
previous stroke (OR = 11.35, 95% CI = 1.62 to 79.73,
p = 0.015) significantly predicted CDR>1 (table 3).
Univariate regression analysis revealed that NIHSS,
performance on MDRS I/P, and ADAS-cog were associated
with BI. Multivariate regression analysis revealed that
only NIHSS (R
2
= 0.096, p = 0.011) contributed signi-
ficantly to the variance of BI (table 4). Univariate
analysis revealed that age, education, NIHSS, pre-stroke
IQCODE, WMC, and performance on all the three psycho-
metric tests were associated with IADL. Multivariate regres-
sion analysis revealed that NIHSS (R
2
= 0.327, p,0.001),
performance on MDRS I/P (R
2
= 0.139, p,0.001), age
(R
2
= 0.052, p = 0.011), and pre-stroke IQCODE (R
2
= 0.034,
p = 0.034) contributed significantly to the variance of IADL
(table 4).
DISCUSSION
In this hospital based study, half of the patients with SSVD
had varying severity of cognitive symptoms 3 months after
SSVD. Pre-stroke cognitive decline and previous stroke
predicted severe cognitive impairment. Stroke severity was
associated with impaired basic ADL. Stroke severity and
performance on MDRS I/P contributed most to an impaired
complex ADL 3 months after SSVD.
Although small vessel disease is the most common cause
for small infarct, other vascular aetiologies are also recog-
nised as causing small infarct.
28 29
In order to include patients
with SSVD, we had rigorously screened out other relevant
vascular aetiologies. In particular, we had looked for relevant
intracranial large artery disease, which is not uncommonly
associated with small infarct in our ethnic group.
16
Indeed, 14
patients (13.2%) with relevant small infarct were found to
have relevant intracranial large artery disease and were thus
excluded. As patients who had relevant small infarcts with
unknown vascular aetiology were also excluded, the subjects
in our study probably represent patients with a homogeneous
aetiology, namely small vessel disease.
In the present study, we attempted to describe cognitive
impairment among patients with SSVD from three different
perspectives: (a) cognitive symptoms as noted by the patients
or their close informants, (b) varying severity of the cognitive
symptoms as graded by the CDR, and (c) performance in
various psychometric tests.
Cognitive symptoms
Cognitive symptoms were noted by more than half (52%) of
the patients 3 months after SSVD. Majority of the patients
(82.1%) claimed that there was a clear temporal relation
between the cognitive symptoms and the index stroke. This is
consistent with most studies showing that small infarcts
Table 2 Comparison of controls and patients of varying CDR grading
Variables Controls CDR = 0 CDR = 0.5 CDR>1
n 42 39 (52%) 26 (34.7%) 10 (13.3%)
Age 69.6 (9.9) 69.4 (12.8) 70.2 (8.8) 77.1 (8.7)
Gender, n (% F) 24 (58.5%) 20 (51.3%) 12 (46.2%) 4 (40.0%)
Years of education 5.4 (4.6) 4.1 (3.7) 5.8 (4.4) 5.5 (4.8)
NIHSS NA 4.1 (2.7) 5.8 (4.4) 5.5 (4.8)
HT NA 32 (82.1%) 23 (88.5%) 9 (90.9%)
DM NA 18 (46.2%) 6 (23.1%) 4 (40.0%)
Previous stroke NA 8 (21.6%) 4 (15.4%) 5 (50.0%)
Hyperlipidaemia NA 11 (28.2%) 9 (34.6%) 3 (30.0%)
Heart diseases NA 5 (12.8%) 4 (15.4%) 1 (10.0%)
History of smoking NA 13 (33.3%) 13 (50.0%) 3 (30.0%)
History of drinking NA 11 (28.2%) 9 (34.6%) 1 (10.0%)
BI 20.0 (0.0)18.5 (3.4) 18.5 (3.8) 16.6 (3.0)
IADL 0.1 (0.2)`10.7 (0.7) 0.8 (0.7) 1.6 (1.0)`1
Sites of recent infarcts
Striatocapsule NA 9 (23.1%) 4 (15.4%) 2 (20.0%)
Cerebral white matter NA 5 (12.8%) 8 (30.8%) 4 (40.0%)
Thalamus NA 10 (25.6%) 7 (26.9%) 1 (10.0%)
Cerebellum NA 0 (0.0%) 1 (3.8%) 0 (0.0%)
Brainstem NA 15 (38.5%) 6 (23.1%) 3 (30.0%)
No. of silent infarcts*NA 0.5 (2.0) 0.0 (2.0) 2.0 (2.3)
WMC*NA 5.5 (8.5) 4.0 (3.0) 13.0 (14.8)
Pre-stroke IQCODE NA 3.1 (0.2)13.3 (0.3)` 3.7 (0.4)`1
MMSE 27.7 (2.1)126.2 (3.4)24.9 (3.8) 19.6 (6.3)`1
MDRS I/P 34.1 (3.0)`130.0 (5.2) 29.0 (5.4) 21.3 (8.3)`1
ADAS-cog 9.5 (3.4)`114.4 (6.9) 16.0 (6.7) 26.9 (16.6)`1
Values are mean (SD) or number (%).
*Values are median (interquartile range); differs significantly from controls, p,0.05; `differs significantly from
CDR = 0, p,0.05; 1differs significantly from CDR = 0.5, p,0.05; differs significantly from CDR>1, p,0.05.
NA, not applicable, F, female.
Table 3 Determinants for CDR>1
Statistical analysis p OR*
95% CI
Lower Upper
Univariate binomial regression
Age NS 1.08 1.00 1.17
Gender NS 0.69 0.18 2.67
Years of education NS 1.04 0.89 1.22
Pre-stroke IQCODE ,0.01 41.59 4.82 358.57
NIHSS 0.05 1.29 1.00 1.66
HT NS 1.64 0.19 14.38
DM NS 1.14 0.29 4.45
Previous stroke 0.04 4.25 1.06 17.06
Hyperlipidemia NS 0.96 0.23 4.12
Heart diseases NS 0.69 0.08 6.13
History of smoking NS 0.64 0.15 2.72
History of alcohol drinking NS 0.25 0.03 2.11
Total no. silent infarcts NS 1.30 0.82 2.05
WMC 0.01 1.26 1.07 1.48
Stepwise multivariate binomial regression
Pre stroke IQCODE ,0.001 95.52 7.37 1238.42
Previous stroke 0.015 11.35 1.62 79.73
NS, not significant.
*Adjusted OR for stepwise multivariate binomial regression.
Cognitive impairment and functional outcome after SSVD 563
www.jnnp.com
located in various subcortical or infratentorial regions are
capable of causing cognitive impairment as well as sensori-
motor deficits.
3 6 7 30–33
As many of these patients already had
varying degrees of pre-stroke cognitive decline, as suggested
by their impaired pre-stroke IQCODE (table 2), the recent
small infarct probably induced only a further stepwise
deterioration of pre-existing cognitive impairment, rather
than inducing a new onset cognitive impairment. The
magnitude of cognitive deterioration that was associated
with the recent small infarct was not ascertained in this
study. Small case series suggest that the magnitude of
cognitive deterioration may vary from subtle
30
to severe.
3
Our attempt to classify the types of cognitive impairment
based on subjective symptoms was only exploratory. Patient
complaint of memory loss may in fact be due to slow thinking
or word finding difficulty; the reverse may also be true.
Therefore, further study is needed to investigate the type of
cognitive impairment by psychometric tests that evaluate
specific cognitive domains.
CDR grading and performance of psychometric tests
To assess the frequency of the varying severity of cognitive
impairment, we arbitrarily divided our patients into three 3
groups based on the CDR grading. To date, there is no
standardised method to define the varying severity of
cognitive impairment associated with cerebrovascular dis-
ease.
34
We did not use cutoff scores of psychometric tests to
grade cognitive impairment because controversy still exists in
the selection of such tests. Furthermore, valid cutoff scores
for defining different severity of cognitive impairment have
not been determined in our ethnic group for this disease
entity. We had chosen CDR simply because of its ease of
administration and because other recent studies had also
applied CDR in grading cognitive impairment associated with
small vessel disease.
25 35
As CDR was originally designed for
grading dementia of the Alzheimer’s type,
20 21
memory bears
greater weight in the rating algorithm. Thus, it may not
reflect the true impact of non-memory cognitive symptoms
that are associated with SSVD. In fact, the three patients with
isolated slow thinking were graded as CDR of 0 according to
the rating algorithm. Therefore, caution should be taken in
interpreting the grading of the patients. However, given this
limitation, we found that the performance in various
psychometric tests of our patients correlated roughly with
the severity of CDR grading (table 2). This correlation lends
support to the validity of this functional based rating scale in
estimating the severity of cognitive impairment in patients
with SSVD. Minor modification by including symptoms of
executive dysfunction such as slow thinking as a primary
category may improve the sensitivity and accuracy of CDR in
grading cognitive impairment in SSVD.
The frequency of CDR >1 or those with greater cognitive
impairment was 13.3%. A few hospital based studies
demonstrate that the frequency of dementia after SSVD
ranges from 5% to 27% at 1 month to 3 years after SSVD
36–40
(table 5). Our frequency of 13.3% fell within this range. The
differences between the studies are probably caused by
differences in the study population, method of defining
dementia or severe cognitive impairment, and follow up time
(table 5). The frequency of severe cognitive impairment in
SSVD is less than that in large arterial stroke (20%–32%).
36 40
We must emphasise that patients with severe cognitive
impairment had only mild stroke severity (mean NIHSS of
5.5). However, they had significantly more impaired func-
tioning in complex ADL than those with lesser cognitive
impairment even though stroke severity and functioning in
basic ADL were similar. These observations suggest that
ignoring the cognitive impact among patients with severe
cognitive impairment may underestimate the consequence of
SSVD if motor deficits and basic ADL only are assessed in this
group of patients.
Patients with CDR of 0.5 were almost identical to patients
with CDR of 0 in terms of basic demographic data, stroke
severity, vascular risk factors, neuroimaging features, post-
stroke functional impairment, and performance in all
psychometric tests. Interestingly, we found that only pre-
stroke cognitive decline was significantly more impaired in
patients with CDR of 0.5 than that in patients with CDR of 0
even though both groups had similar performance in
psychometric tests. There are several possibilities that may
explain the subtle difference between these two groups.
Firstly, as addressed above, the CDR grading may undergrade
the cognitive consequence of patients with mild, non-
memory cognitive symptoms. Including these patients into
CDR of 0 may mask any genuine differences between the two
groups. Secondly, the various psychometric tests that were
used in our study may not be sensitive enough to reveal a
significant difference between these two groups of patients.
Thirdly, the pre-stroke IQCODE may be more sensitive in
detecting mild post-stroke cognitive impairment in patients
with SSVD than other psychometric tests that were used in
our study. Fourthly, the sample size may not be powerful
enough to demonstrate a statistical difference in the
performance of psychometric tests between the two groups.
The mean scores of all psychometric tests for patients with
CDR of 0.5 appeared to be slightly worse than that for
patients with CDR of 0. A larger sample size may possibly
reveal a statistical difference in the performance of psycho-
metric tests between these two groups of patients. Lastly, any
combinations of the above may be responsible.
Apart from the three patients who had mild isolated slow
thinking, all the remaining patients with CDR of 0 denied any
cognitive symptoms before or after the stroke. Their
performances in MDRS I/P and ADAS-cog were significantly
worse than that of controls even though both groups had
similar basic demographic data. We postulate that this
difference is probably due to the small infarcts and WMC.
Studies demonstrate that both small subcortical infarcts and
WMC are associated with impaired performance in psycho-
metric tests among healthy individuals who are free of stroke
or cognitive symptoms.
541
We did not perform neuroimaging
for our controls. Silent infarcts and age related WMC might
Table 4 Association of clinical, cognitive, and
radiological features with functional outcome
Statistical analysis
BI IADL
R
2
p (F change) R
2
p (F change)
Univariate linear regression
Age 0.003 NS 0.265 ,0.001
Education 0.004 NS 0.142 ,0.001
NIHSS 0.084 0.015 0.278 ,0.001
Pre-stroke IQCODE 0.04 NS 0.236 ,0.001
MMSE 0.036 NS 0.278 ,0.001
MDRS I/P 0.098 0.009 0.285 ,0.001
ADAS-cog 0.072 0.026 0.244 ,0.001
Previous stroke 0.013 NS 0.001 NS
Heart disease 0.005 NS ,0.001 NS
No. of silent infarcts 0.005 NS 0.035 NS
WMC 0.007 NS 0.177 ,0.001
Stepwise multivariate linear regression
NIHSS 0.096 0.011 0.327 ,0.001
MDRS I/P NA NA 0.139 ,0.001
Age NA NA 0.052 0.011
Pre-stroke IQCODE NA NA 0.034 0.034
Total variance
explained
9.60% 55.20%
NS, not significant
564 Mok,Wong,Lam,etal
www.jnnp.com
also be present among our controls. Furthermore, other non-
vascular factors may be affecting the cognition of patients
with SSVD who have no cognitive symptoms. Hence, further
study is needed to evaluate the relation between ischaemic
lesions and cognitive impairment both in controls and in
patients with SSVD who have no cognitive symptoms.
On the whole, patients with CDR of 0 or 0.5 had mild
stroke severity and minimal functional impairment. The
immediate relevance of detecting patients with mildly
disabling cognitive symptoms or mild impairment in the
performance of psychometric tests but without cognitive
symptoms is uncertain. However, as cognitive impairment in
small vessel disease has a propensity to progress,
910
it is of
paramount importance to study any potentially reversible
risk factors that are associated with further cognitive
deterioration among this group of patients.
We had excluded 17 patients with relevant small infarcts
but who had unknown vascular aetiology or who had not
participated in the cognitive assessment. As these patients
were significantly older than the analysed subjects, we may
have underestimated the frequency of cognitive impairment
in our present study. Furthermore, patients with major
depression were also among those who were excluded
(n = 3). It is currently believed that subcortical lesions,
cognitive impairment, and depression are all inter-related.
42
Hence, our estimation on the frequency of the varying
severity of cognitive impairment was only conservative.
Determinants for CDR >1
Pre-stroke cognitive decline and previous stroke were
identified to predict severe cognitive impairment 3 months
after SSVD. The cause for pre-stroke cognitive decline has not
been investigated in the present study. Understanding this
cause is important to the management of post-stroke
cognitive impairment, as the disease process that has been
affecting the patients pre-stroke is probably continuing to
affect them post-stroke. Studies on pre-stroke cognitive
decline among patients with stroke in general show that
both ischaemic and degenerative pathology predict pre-stroke
cognitive decline.
43–45
Further study is needed to explore the
aetiology and determinants of pre-stroke cognitive decline
among patients with SSVD. It is interesting to find that
previous stroke, rather than ischaemic lesions (WMC, small
infarcts) predictes post-stroke severe cognitive impairment in
SSVD. Other variables that have not been explored in the
present study may be associated with previous stroke and
with post-stroke cognitive impairment. A recent study
suggests that atrophy of the cortex and hippocampus, which
can only be partially explained by WMC or Alzheimer’s
pathology, was associated with dementia in patients with
subcortical lacunes who were recruited randomly from a
dementia clinic.
46
It may be worthwhile to explore the
relation between previous stroke, atrophy of the cortex or
hippocampus, and post-stroke severe cognitive impairment in
SSVD. We must emphasise that caution should be taken in
interpreting the results from the regression analysis. As our
sample size was small, the confidence intervals of the two
variables identified as significant were wide (table 3). A
larger study is needed to confirm the importance of pre-
stroke cognitive decline and previous stroke in predicting
post-stroke severe cognitive impairment in SSVD.
Functional outcome
Multivariate regression analysis demonstrated that perform-
ance in basic ADL as measured by BI was significantly
associated with stroke severity (NIHSS) in our present
cohort. This is consistent with other studies of SSVD showing
that motor impairment significantly predicts physical depen-
dence.
15 39
Other variables that have been shown to predict
physical dependence in SSVD are WMC, age, diabetes,
previous stroke, and type of lacunar syndrome.
15 39
However, the influence of cognitive impairment on the
performance of complex ADL in SSVD has rarely been
studied. We have demonstrated that stroke severity and
performance in MDRS I/P contributes most to the variance in
IADL. Age and pre-stroke IQCODE has a lesser contribution
to its variance. As MDRS I/P evaluates executive dysfunction,
our findings suggest that executive dysfunction predicts
impairment in complex ADL among patients with SSVD.
Tests that are biased towards evaluation of memory and
language impairment, namely MMSE and ADAS-cog, were
found not to be associated with performance of complex ADL
in our subjects. A recent study also suggests that executive
dysfunction among stroke patients in general is associated
with reduced performance in both basic and complex ADL.
53
Our findings highlight the importance of cognitive impair-
ment, and in particular, executive dysfunction, in affecting
the functional outcome of patients with SSVD.
Authors’ affiliations
.....................
V C T Mok, A Wong, W W M Lam, Y H Fan, W K Tang, T Kwok,
A C F Hui, K S Wong, The Chinese University of Hong Kong, Hong
Kong, China
Competing interests: none declared
REFERENCES
1Erkinjuntti T, Inzitari D, Pantoni L, et al. Research criteria for subcortical
vascular dementia in clinical trials. J Neural Transm 2000(Suppl);59:23–30.
2Roman GC, Erkinjuntti T, Wallin A, et al. Subcortical ischemic vascular
dementia. Lancet Neurol 2002;1:426–36.
3Tatemichi TK, Desmond DW, Prohovnik I, et al. Confusion and memory loss
from capsular genu infarction:a thalamcortical disconnection syndrome?
Neurology 1992;42:1966–79.
4Cummings JL. Frontal-subcortical circuits and human behavior. Arch Neurol
1993;50:873–80.
Table 5 Frequency of severe cognitive impairment / dementia after SSVD
Studies n Study population Dementia criteria
Follow up
duration
Frequency
(%)
Tatemichi
36
227 With/without previous stroke
or pre-stroke dementia
Clinical diagnosis 2 years 11
Wolfe
37
11 Multiple lacunes, with/without
pre-stroke dementia
Clinical diagnosis .1 month 27
Miyao
38
215 First stroke, no pre-stroke
dementia
DSM-III-R & MMSE
,20
2 years 11.2
Samuelsson
39
81 First stroke Clinical diagnosis 3 years 11
Desmond
40
170 With/without previous stroke
or pre-stroke dementia
DSM-III-R 3 months 25.9
Present study 75 With/without previous stroke
or pre-stroke dementia
CDR>1 3 months 13.3
DSM-III-R, Diagnostic and Statistical Manual of Mental Disorders, 3rd edition, revised.
Cognitive impairment and functional outcome after SSVD 565
www.jnnp.com
5Kramer JH, Reed BR, Mungas D, et al. Executive dysfunction in subcortical
ischemic vascular disease. J Neurol Neurosurg Psychiatry 2002;72:217–20.
6Naeser MA, Alexander MP, Helm-Estabrooks, et al. Aphasia with
predominantly subcortical lesion sites. Arch Neurol 1982;39:2–14.
7Giroud M, Lemesle M, Madinier G, et al. Unilateral lenticular infarcts:
radiological and clinical syndromes, aetiology, and prognosis. J Neurol
Neurosurg Psychiatry 1997;63:611–15.
8Babikian V, Ropper AH. Binswanger’s disease: a review. Stroke
1987;18:2–12.
9Meyer JS, Xu G, Thornby J, et al. Is mild cognitive impairment prodromal for
vascular dementia like Alzheimer’s disease?Stroke 2002;33:1981–5.
10 Aharon-Peretz J, Daskovski E, Mashiach T, et al. Natural history of dementia
associated with lacunar infarctions. J Neurol Sci 2002;203–204:53–5.
11 Korczyn AD. Mixed dementia–the most common cause of dementia. Ann NY
Acad Sci 2002;977:129–34.
12 Bamford J, Sandercock P, Jones L, et al. The natural history of lacunar
infarction: The Oxfordshire Community Stroke Project. Stroke
1987;18:545–51.
13 Sacco SE, Whisnant JP, Broderick JP, et al. Epidemiological characteristics of
lacunar infarcts in a population. Stroke 1991;22:1236–41.
14 Rothrock JF, Lyden PD, Brody ML, et al. An analysis of ischemic stroke
in an urban Southern California population. Arch Intern Med
1993;153:619–24.
15 Clavier I, Hommel M, Besson G, et al. Long term prognosis of symptomatic
lacunar infarcts–a hospital-based study. Stroke 1994;25:2005–9.
16 Wong KS, Gao S, Chan YL, Ringelstein EB, et al. Mechanisms of acute
cerebral infarctions in patients with middle cerebral artery stenosis: A
diffusion-weighted imaging and microemboli monitoring study. Ann Neurol
2002;52:74–81.
17 Wong KS, Li H, Chan YL, et al. Use of transcranial Doppler ultrasound to
predict outcome in patients with intracranial large-artery occlusive disease.
Stroke 2000;31:2641–7.
18 American Psychiatric Association.Diagnostic and statistical manual
disorders, 4th ed. (DSM-IV). Washington, DC: American Psychiatric
Association, 1994.
19 Wahlund LO, Barkhof F, Fazekas F, et al. on behalf of the European Task
Force on Age Related White Matter Changes. A new rating scale for age-
related white matter changes applicable to MRI and CT. Stroke
2001;32:1318–22.
20 Huges CP, Berg L, Danziger L, et al. A new clinical scale for the staging of
dementia. Br J Psychiatry 1982;140:566–72.
21 Morris JC. The clinical dementia rating (CDR): current version and scoring
rules. Neurology 1993;43:2412–14.
22 Chiu HFK, Kee HC, Chung WS, et al. Reliability and validity of the Cantonese
version of Mini-mental state examination–a preliminary study. J Hong Kong
Coll Psychiatry 1994;4(Suppl 2):25–8.
23 Chu LW, Chiu KC, Hui SL, et al. The reliability and validity of the Alzheimer’s
disease assessment scale cognitive subscale (ADAS-cog) among the elderly
Chinese in Hong Kong. Ann Acad Med Singapore 2000;29:474–85.
24 Chan AS, Choi MK, Salmon DP. The effects of age, education, and gender on
the Mattis dementia rating scale performance of elderly Chinese and
American individuals. J Gerontol Psychol Sci 2001;56B:356–63.
25 Mungas D, Jagust WJ, Reed BR, et al. MRI predictors of cognition in
subcortical ischemic vascular disease and Alzheimer’s disease. Neurology
2001;57:2229–35.
26 Fuh JL, Teng EL, Lin KN, et al. The Informant Questionnaire on Cognitive
Decline in the Elderly (IQCODE) as a screening tool for dementia for a
predominantly illiterate Chinese population. Neurology 1995;45:92–6.
27 Lawton MP, Brody EM. Assessment of older people: self-maintaining and
instrumental activities of daily living. Gerontologist 1969;9:179–86.
28 Bamford JM, Warlow CP. Evolution and testing of the lacunar hypothesis.
Stroke 1988;19:1074–82.
29 Baumgartner RW, Sidler C, Mosso M, et al. Ischemic lacunar stroke in
patients with and without potential mechanism other than small-artery disease.
Stroke 2003;34:653–9.
30 Van Zandvoort MJE, Kappelle LJ, Algra A, et al. Decreased capacity for
mental effort after single supratentorial lacunar infarct may affect performance
in everyday life. J Neurol Neurosurg Psychiatry 1998;65:697–702.
31 Caplan LR. Caudate infarcts. In: Donnan G, Norrving B, Bamford J,
Bogousslavsky J, eds. Subcortical stroke, 2nd edn. New York: Oxford
University Press, 2002:209–23.
32 de Freitas GR, Bogousslavsky J. Thalamic infarcts. In: Donnan G, Norrving B,
Bamford J, Bogousslavsky J, eds. Subcortical stroke, 2nd edn. New York:
Oxford University Press, 2002:255–85.
33 Malm J, Kristensen B, Karlsson T, et al. Cognitive impairment in young adults
with infratentorial infarcts. Neurology 1998;51:433–40.
34 Bowler JV. The concept of vascular cognitive impairment. J Neurol Sci
2002;203:11–15.
35 Mungas D, Reed BR, Jagust WJ, et al. Volumetric MRI predicts rate of
cognitive decline related to AD and cerebrovascular disease. Neurology
2002;59:867–73.
36 Tatemichi TK, Foulkes MA, Mohr JP, et al. Dementia in stroke survivors in the
stroke data bank cohort, prevalence, incidence, risk factors, and computed
tomographic findings. Stroke 1990;21:858–66.
37 Wolfe N, Linn R, Babikian VL, et al. Frontal systems impairment following
multiple lacunar infarcts. Arch Neurol 1990;47:129–32.
38 Miyao S, Takano A, Teramoto J, et al. Leukoaraiosis in relation to prognosis
for patients with lacunar infarction. Stroke 1992;23:1434–8.
39 Samuelsson M, Soderfeldt B, Olsson GB. Functional outcome in patients with
lacunar infarction. Stroke 1996;27:842–6.
40 Desmond DW, Moroney JT, Paik MC, et al. Frequency and clinical
determinants of dementia after ischemic stroke. Neurology
2000;54:1124–31.
41 Schmidt R, Fazekas F, Offenbacher H, et al. Neuropsychologic correlates of
MRI white matter hyperintensities: a study of 150 normal volunteers.
Neurology 1993;43:2490–4.
42 Carota A, Staub F, Bogousslavsky J. Emotions, behaviours and mood changes
in stroke. Curr Opin Neurol 2002;15:57–69.
43 Henon H, Pasquier F, Durieu I, et al. Preexisting dementia in stroke patients.
Baseline frequency, associated factors, and outcome. Stroke
1997;28:2429–36.
44 Pohjasvaara T, Mantyla R, Aronen HJ, et al. Clinical and radiological
determinants of prestroke cognitive decline in a stroke cohort. J Neurol
Neurosurg Psychiatry 1999;67:742–8.
45 Barba R, Castro MD, del Mar Morin M, et al. Prestroke dementia.
Cerebrovasc Dis 2001;11:216–24.
46 Fein G, Di Sclafani V, Tanabe J, et al. Hippocampal and cortical atrophy
predict dementia in subcortical ischemic vascular disease. Neurology
2000;55:1626–35.
47 Pohjasvaara T, Leskela M, Vataja R, et al. Post-stroke depression, executive
dysfunction and functional outcome. Eur J Neurol 2002;9:269–75.
566 Mok,Wong,Lam,etal
www.jnnp.com
