REVIEW
www.nature.com/clinicalpractice/neuro
Vascular cognitive impairment
Ola A Selnes* and Harry V Vinters
S U M M A RY
Cognitive impairment commonly accompanies clinical syndromes
associated with vascular disease of the brain. Because of evolving
definitional criteria, however, the frequency of cognitive impairment
attributable to cerebrovascular disease is difficult to determine. Dementia
occurs in up to one-third of elderly patients with stroke, a subset of whom
have Alzheimers disease (AD) rather than a pure vascular dementia
syndrome. In fact, pure vascular dementia has been shown to be uncommon
in most large autopsy series. A mixed etiology of AD and cerebrovascular
disease is thought to become more common with increasing age, although
no clinical criteria for the diagnosis of AD with cerebrovascular disease are
currently available. Epidemiological studies have implicated subcortical
small-vessel disease as a risk factor for cognitive impairment and dementia,
but the cognitive expression and clinical significance of MRI white matter
changes in individual patients is difficult to establish. The frequency
of specific neuropathologic features of vascular cognitive impairment
depends largely on study inclusion criteria. Cerebral meningocortical
microangiopathies with distinctive clinicopathological profiles are
associated with dementia in both sporadic cases and familial syndromes.
In patients with AD, the contribution of amyloid- protein to the degree of
cognitive impairment has not been clearly defined.
KEYWORDS brain ischemia, CADASIL, cerebral amyloid angiopathy,
cerebrovascular disorders, dementia
REVIEW CRITERIA
We searched PubMed (19902005) for articles, including electronic early
release articles. Search terms included dementia, CADASIL, Alzheimers
disease, cerebrovascular disorders, cerebral infarction, brain ischemia,
cerebral amyloid angiopathy, cognition disorders, cognition disorders/
pathology, neuropsychology, dementia/pathology and brain/pathology.
We prioritized articles according to several criteria, including large study sample
sizes, inclusion of a control group, and assessment of cognition by standardized
neuropsychological tests.
OA Selnes is a professor in the Cognitive Neuroscience Division, Department
of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD,
USA, with a joint appointment in the Department of Psychiatry. HV Vinters
holds the Daljit S. and Elaine Sarkaria Chair in Diagnostic Medicine, and
is Professor of Pathology and Laboratory Medicine, and Neurology, at the
University of California, Los Angeles, CA, USA.
Correspondence
*Division of Cognitive Neuroscience, Johns Hopkins University School of Medicine,
Reed Hall East 2, 1620 McElderry Street, Baltimore, MD 21205-1910, USA
oselnes@jhmi.edu
Received 13 March 2006 Accepted 21 July 2006
www.nature.com/clinicalpractice
doi:10.1038/ncpneuro0294
538 NATURE CLINICAL PRACTICE NEUROLOGY
2006 Nature Publishing Group
INTRODUCTION
Cognitive impairment is part of the clinical
presentation of several conditions associated
with cerebrovascular disease, but the precise
frequency of vascular cognitive disorders is diffi-
cult to ascertain. Clinicopathological studies of
various selected or unselected populations have
shown a remarkable heterogeneity of vascular
and brain parenchymal lesions,13 and classifica-
tion schemes based on clinical,4 radiological5 and
neuropathologic1,6 criteria have been proposed.
In this Review, only the major subtypes of vascular
cognitive impairment will be considered.4
Diagnostic criteria for vascular dementia
have evolved over the past two decades,7 but
several obstacles to widespread acceptance of
these definitions are recognized. Modeled after
Alzheimers disease (AD), early definitions of
vascular dementia (VaD) disproportionately
emphasized deficits of new learning and
memory, rather than the pattern of motor
slowing and executive deficits typically asso-
ciated with VaD. The requirement that patients
with VaD should have cognitive impairment
of sufficient severity to interfere with activi-
ties of daily living is now acknowledged as too
restrictive, because it does not allow cases to be
identified early enough for diagnostic inter-
ventions. The severity and type of lesions on
neuroimaging required for a diagnosis of VaD
remain controversial, and the requirement of
a temporal link between the onset of cognitive
changes and neuroimaging changes restricts
case identification predominantly to VaD
secondary to infarcts.8
The term vascular cognitive disorder was
proposed by Sachdev9 to define vascular cogni-
tive deficits of sufficient severity to meet criteria
for a diagnosable disorder. It was intended as an
umbrella term to include the spectrum of impair-
ment from mild vascular cognitive impairment
(VCI) to VaD.10 The development of a single,
uniform set of criteria that apply to all subtypes
of VCI has clearly been challenging, and some
experts have recommended that separate criteria
OCTOBER 2006 VOL 2 NO 10

be developed for certain subtypes, such as non-
infarct-related subcortical small-vessel disease.11
The aim of this Review is to provide a brief over-
view of the current state of scientific knowledge
in the field of VCI, with an emphasis on its
neurocognitive and neuropathologic aspects.
SUBTYPES OF VASCULAR COGNITIVE
IMPAIRMENT
Post-stroke dementia
The prognosis for recovery of cognitive symp-
toms, including aphasia, after an initial stroke
is generally favorable, but some patients do not
show the expected recovery, and instead develop
persistent or progressive cognitive decline.12
Because older patients at risk for stroke have an
increased risk of dementia even in the absence
of a stroke, the inclusion of elderly controls is
crucial for interpreting the results of post-stroke
dementia (PSD) studies.13,14 PSD has long been
considered the prototypical subtype of VaD, and
it might therefore be expected that all patients
with PSD would meet the criteria for VaD.
Surprisingly, however, only one-third of such
patients have been reported to meet the diag-
nostic criteria for AD.12,15 This finding is consis-
tent with epidemiological data showing that risk
factors for cerebrovascular disease are common in
patients with AD,16 and also that many patients
diagnosed with PSD have cognitive impairment
even before their stroke.12,17
The most important demographic predictor
of PSD is age; the association with stroke risk
factors is less robust.12 Degree of pre-existing
subcortical white matter disease, infarct
volume, and global and medial temporal lobe
atrophy, have been identified as some of the
relevant imaging determinants of post-stroke
dementia,18 and cortical hypoperfusion might
also play an important role.14 Stroke character-
istics, such as lesion volume and location, appear
to be less predictive of PSD. A greater degree of
severity of cognitive impairment after stroke has
been associated with increased risk of PSD.15,19
The apolipoprotein E 4 allele is a risk factor for
AD, but it does not appear to be associated with
increased risk of PSD.20
Strategic infarct dementia
The view that a certain threshold volume of
brain tissue loss (e.g. >50100 ml) predictably
causes dementia is no longer widely accepted;
OCTOBER 2006 VOL 2 NO 10 SELNES AND VINTERS
2006 Nature Publishing Group
REVIEW
www.nature.com/clinicalpractice/neuro
smaller infarcts in particular regions, for
example in the deep central gray matter, might
have an equally important role in causing
dementia.1 Lacunar infarcts involving the
thalamus, internal capsule and basal ganglia are
sometimes associated with surprisingly wide-
spread cognitive effects, including confusion
and memory impairment.21 Infarcts involving
the dorsomedial and anterior thalamus might
also produce significant executive symptoms
and profound amnesia, which can persist in
some cases.22 Prospective follow-up studies
have shown that, although initially quite severe,
cognitive symptoms associated with strategic
infarcts are often reversible by 12 months,
and they are therefore not a common cause of
persistent dementia.23
In summary, there is substantial evidence that
stroke in the context of advanced age confers an
increased risk of dementia. Nonetheless, the risk
of PSD appears to be more closely related to the
severity of pre-existing white matter abnormali-
ties, atrophy and hemodynamic factors than to
the stroke characteristics themselves. Although
generally easily recognized clinically, PSD repre-
sents only the proverbial tip of the iceberg of
the spectrum of VCI.24
Alzheimers disease with cerebrovascular
disease
Determining the neurobehavioral and neuro-
imaging correlates of ischemic brain lesions
occurring in the context of significant AD altera-
tions can be a daunting task, and the usefulness
of the traditional strict dichotomization between
AD and VaD has recently been challenged.25
Evidence is accumulating that AD is commonly
associated with vascular risk factors, including
diabetes,26 hypertension and smoking.16 Whether
these associations are causal or coincidental
co-occurrences of common age-associated
conditions is not known.
Many individuals with AD, especially those
beyond 85 years of age, show significant
vascular comorbidity, to the extent that they are
more accurately characterized as having mixed
vascularAD dementia. 27 In longitudinal
studies and large clinicopathological series
with necropsy confirmation, the observed
morpho-anatomical substrates are largely
determined by study entry criteria. Autopsy
investigations in individuals who develop
dementia after large cortical or subcortical
NATURE CLINICAL PRACTICE NEUROLOGY 539
REVIEW
www.nature.com/clinicalpractice/neuro
cystic infarcts will show different patterns
of lesions from studies in which patients are
selected for a high likelihood of deep lacunar
infarcts or subcortical white matter alterations,
including leukoaraiosis.6,28 Such populations
of patients are also likely to differ in their
clinical progression; those with small deep
lacunar infarcts (especially within the white
matter) or multiple microinfarcts are less
likely to show a stepwise progression of cogni-
tive impairment than are those with regions
of substantial cystic encephalomalacia. In one
large autopsy series, pure VaD was seen in 9.4%
of 900 individuals with dementia, but in only
2.9% of patients with the clinical diagnosis of
probable or possible AD.29 Some investigators
suggest that VaD might account for only 23%
of dementias,30 and others claim that pure VaD
is almost nonexistent in large dementia series,
even when careful necropsies are performed.31
Even though a mixed etiology is likely to be
more common than either pure AD or VaD
among older patients, there are no current
clinical criteria for ante-mortem diagnosis
of mixed dementia.4
Subcortical small-vessel ischemic disease
White matter hyperintensities and lacunar
infarcts demonstrated by MRI are generally
considered to be evidence of small-vessel or
microvascular ischemic disease, although some
are related to dilated perivascular spaces.1 White
matter abnormalities on MRI are common in
otherwise healthy community-dwelling indi-
viduals.32 The prevalence of these abnormali-
ties increases sharply with age, and other risk
factors include hypertension, diabetes and
genetic factors.33,34 In patients with known
cardiovascular disease, such as candidates for
coronary artery bypass grafting, the preva-
lence has been reported to be as high as 50%.35
Pathologically, MRI white matter signal abnor-
malities reflect focal and diffuse lesions of the
subcortical and periventricular white matter, as
well as lacunes and microinfarcts of the central
gray matter.36
Because white matter abnormalities are
so common in otherwise asymptomatic
individuals, they are often considered to be
benign. There is increasing evidence, however,
that white matter abnormalities on MRI are
associated with an increased risk of dementia.
In cross-sectional studies, individuals with
540 NATURE CLINICAL PRACTICE NEUROLOGY
2006 Nature Publishing Group
more-severe white matter lesions had a twofold
increased risk of dementia.37 In prospective
studies, the presence of periventricular white
matter lesions at baseline was found to double
the risk of future dementia.21 There is also
evidence that progression of white matter
disease is accompanied by a decline in cogni-
tive performance, supporting an etiological
link between the white matter changes and
cognitive decline.32,38
Although epidemiological studies have clearly
established an association of subcortical white
matter lesions with poorer performance in
specific cognitive domains,39 the cognitive
expression of subcortical small vessel disease in
individual patients is highly variable. In a patient
who presents with mild cognitive impairment
and mild subcortical white matter lesions, the
etiological significance of the white matter
lesions cannot be easily determined. Neither a
history of risk factors for cerebrovascular disease
nor the clinical presentation can establish that
the cognitive symptoms are causally related to
the white matter findings. On the one hand,
there are case reports of patients with MRI-
documented widespread white matter disease
whose cognitive functioning is normal even
by detailed neuropsychological testing.40 On
the other hand, it is also well known that white
matter abnormalities are commonly seen in
patients with AD.41
Although there is no cognitive profile that is
entirely specific for subcortical vascular disease,
neuropsychological testing can nevertheless
identify whether the overall pattern is predomi-
nantly cortical or subcortical. In patients who
present with cognitive changes in the context
of risk factors for cerebrovascular disease,
and with MRI findings of subcortical white
matter disease, the finding of a predominantly
subcortical profile will significantly increase
the likelihood that the cognitive symptoms
are of vascular origin. The profile of neuro-
psychological test findings alone cannot rule
out possible primary or coexisting AD, but if
follow-up testing demonstrates stable or rela-
tively minor change over time, this can further
support a diagnosis of VCI.
An important but unresolved question is
whether subcortical small-vessel disease by
itself can lead to cognitive impairment of
sufficient severity to meet the criteria for
dementia. Overall, the cognitive expression of
isolated white matter hyperintensities appears
SELNES AND VINTERS OCTOBER 2006 VOL 2 NO 10

to be limited to mild motor and psychomotor
slowing rather than a pervasive, severe degree
of cognitive impairment.42 There is growing
evidence, however, that the subcortical white
matter changes visualized on conventional
MRI might be incomplete markers of more-
widespread, covert hypoxiaischemia-related
injury.43,44 With improvements in the ability
to quantify the overall burden of vascular
disease affecting the brain, including the pres-
ence of hypoperfusion,45,46 microinfarcts47,48
and amyloid angiopathy,49 a more-precise
assessment of the relationship between cerebro-
vascular changes and cognitive impairment
might become possible.
Cerebral autosomal dominant
arteriopathy with subcortical infarcts
and leukoencephalopathy
Cerebral autosomal dominant arteriopathy with
subcortical infarcts and leukoencephalopathy
(CADASIL) is a hereditary non-amyloid type of
small-vessel disease that is commonly associated
with cognitive impairment or dementia.50,51
It results from either a gain or a loss of one
or three cysteine residues in the extracellular
N-terminal region of the NOTCH3 gene, which
is located on chromosome 19p13.52 Affected
deep cerebral arteries show destruction of the
smooth cell layer in their media, accompanied
by progressive wall thickening, and luminal
narrowing caused by this thickening and
fibrosis. Clinically, the condition is character-
ized by recurrent subcortical strokes in patients
aged between 40 years and 60 years. The clinical
manifestations include cognitive impairment
and psychiatric symptoms, with a generally
variable clinical course.53 Some patients may
initially present with only migraine headaches
and no cognitive symptoms. The diagnosis is
made by a combination of characteristic MRI
findings and genetic testing.
NEUROCOGNITIVE CONSIDERATIONS
There has been considerable debate with respect
to the specificity of the cognitive profile asso-
ciated with VCI. Some people have argued that
there is significant overlap between the cogni-
tive profiles of subcortical vascular disease and
cortical dementias such as AD.54,55 Others have
suggested that the cognitive profile of subcortical
vascular disease can be distinguished from that
OCTOBER 2006 VOL 2 NO 10 SELNES AND VINTERS
REVIEW
www.nature.com/clinicalpractice/neuro
of AD principally by milder memory impair-
ment but more-pronounced impairment of
executive functions.56,57
For several reasons, it has been difficult to
resolve the question of the specificity of the
cognitive profile of VCI. First, if all etiologies
of VCI are considered together, there is little
reason to expect a consistent cognitive profile.
The cognitive profile of dementia secondary
to large cortical infarcts might depend on
both the location and volume of the lesion,
and would consequently be highly variable.
Second, because comprehensive cognitive
testing cannot easily be performed in large
clinical or community samples, many of the
studies with detailed testing have been limited
by modest sample sizes.58,59 A third reason why
some investigators have not found consistent
differences between the cognitive profiles of
vascular and AD-type cognitive impairment
is that not all cognitive domains have been
explored in sufficient detail. One of the more
robust neuropsychological characteristics of
patients with subcortical disease is a slowing
of motor and psychomotor speed,39 a feature
that is not typically associated with cortical
dementias. Because most studies have not
included measures of motor speed in their
test battery, however, psychomotor speed has
not generally been considered to be a unifying
feature of VCI.55,59
Studies that have focused on patients with
VCI of sufficient severity to meet the criteria
for dementia have generally found impairments
across most cognitive domains, with more
similarities to than differences from AD. These
studies are likely to have included a subset of
patients with dementia of mixed etiology.55,60
In patients with less-severe cognitive impair-
ment, a more distinctive cognitive profile might
be discernable. Garrett and colleagues reported
that the neuropsychological performance of a
small group of patients with cognitive impair-
ment (but no dementia) was characterized by
disproportionate executive dysfunction and
deficits in verbal retrieval.59 Data from Looi
and colleagues61 also confirm an overall profile
of less-severe memory impairment and greater
executive impairment in VaD. Several inves-
tigators have reported a profile of memory
impairment that includes better preservation
of recognition memory performance relative to
free recall in patients with VCI.58,62 In studies
of patients with CADASIL, who are generally
NATURE CLINICAL PRACTICE NEUROLOGY 541
2006 Nature Publishing Group
REVIEW
www.nature.com/clinicalpractice/neuro
Ischemic vascular dementia (IVD)
Multi-infarct dementia (MID)
Microglial/astroglial proliferation
Secreted factors
Blood vessel-derived factors
Pathogenesis of Consequences for CNS
vascular disease parenchyma
macro/micro [?] (ischemia partial/complete)
Atheroma AS/LH Synapse and dendritic spine loss
CAA (spor/fam) Wallerian degeneration
FMD
?Vasculitis CADASIL Trans-synaptic degeneration
Cardiogenic emboli Others Retrograde cortical neuronal changes
Others
Systematic factors
(hypotension, hypoxia)
Figure 1 Conceptual framework for etiology and pathogenesis of
ischemic vascular dementia, taking into account cerebrovascular
disease, systemic factors and ischemic necrosis of the brain, as well as
retrograde or downstream effects of focal ischemic lesions. Abbreviations:
AS, arteriosclerosis; CAA, cerebral amyloid angiopathy; CADASIL,
cerebral autosomal dominant arteriopathy with subcortical infarcts and
leukoencephalopathy; FMD, fibromuscular dysplasia; LH, lipohyalinosis;
spor/fam, sporadic/familial.
younger and therefore less likely to have
concomitant AD pathologic changes, speed of
processing has consistently been identified as
impaired, with somewhat less-pronounced but
significant deficits in areas of executive perfor-
mance and attention. It has been suggested
that this pattern of impairment represents the
core of the cognitive syndrome associated with
small-vessel subcortical ischemic disease.6365
Only a few studies have examined the cogni-
tive profile of autopsy-confirmed cases of VCI
and dementia.66 These studies have emphasized
considerable variability in the clinical presenta-
tion and the profile of neuropsychological
test findings.
Although the cognitive profile of subcortical
VCI might vary somewhat according to both
etiology and severity of the disease, the pattern
of relative preservation and impairment is
nonetheless helpful in a clinical setting. In a
542 NATURE CLINICAL PRACTICE NEUROLOGY
2006 Nature Publishing Group
patient in whom the history, imaging findings
and clinical presentation are otherwise consis-
tent with subcortical vascular disease, a neuro-
cognitive profile of relatively preserved language
and recognition memory, but with significant
motor and psychomotor slowing and impaired
executive performance, can be considered to
be supportive of a diagnosis of probable VCI
rather than AD.
NEUROPATHOLOGIC SUBSTRATES
OF VASCULAR COGNITIVE
IMPAIRMENT
From a neuropathologic perspective, the
problem of studying ischemic VaD can be
reducedwith the understanding that this is
a huge oversimplificationto examining three
major inter-related, yet functionally separable,
pathophysiologic components: cerebrovascular
disease, systemic mediators of ischemic brain
necrosis (which of course might interact with
cerebrovascular disease), and the CNS paren-
chymal lesions that we recognize as being the
result of cerebrovascular disease and systemic
factors. Once irreversible parenchymal injury
has occurred, there are downstream and
retrograde effects in the CNS that result from
the insultas a function of Wallerian, trans-
synaptic and other types of degeneration
(Figure 1)which almost certainly affect subse-
quent neurobehavioral morbidity in ways that
we do not yet understand.4,67 Subcortical axonal
injury and loss might be key elements in VaD
pathogenesis and progression.68 The identity of
the neuropathologic substrates of leukoaraiosis,
which also affects subcortical white matter, is
still controversial, but might include apoptosis
of oligodendroglia.69
Ischemic parenchymal lesions in patients
with VaD vary in size. In a longitudinal
California-wide study of patients pre-selected
for a high likelihood of subcortical ischemic
lesions, especially lacunar infarcts (based on
ante-mortem neuroimaging studies), we have
arbitrarily subcategorized infarcts seen at
autopsy as cystic infarcts (larger than 1.0 cm in
greatest dimension), lacunar infarcts (grossly
visible on cut sections of the brain but smaller
than 1.0 cm in size), and microinfarcts (not
seen on gross inspection of the cut brain,
but identified by microscopy; see Figure 2).6
Immunohistochemical stains that are especially
helpful for demonstrating microinfarcts include
SELNES AND VINTERS OCTOBER 2006 VOL 2 NO 10
 REVIEW
www.nature.com/clinicalpractice/neuro

those using primary antibodies against CD68 A B

or other macrophagemicroglial markers, and
those using primary antibodies against glial
fibrillary acidic protein (GFAP). Both of these
stains highlight areas of localized proliferation
of cells that react to irreversible ischemic injury.
The roles of astrocytes and microgliaand the
factors that they secretein the progression
of subinfarctive ischemia are not yet known.
Abundant old cortical microinfarcts render C D
the cortical surface irregular; an affected brain
is sometimes described as showing granular
atrophy, even though the causal lesions are
infarctive rather than atrophic.
Cystic infarcts often result from occlusion
of large meningeal arteries affected by athero-
sclerosis, sometimes as a result of atheroemboli
that lodge within themsuch atheroemboli might Figure 2 Microscopic appearance of brain lesions commonly encountered in
also originate in severely atherosclerotic carotid the brains of individuals with cerebrovascular disease and dementia. (A,B) Two
small arteries with marked onion skin-type thickening in the brain of a 68-year-
or vertebral arteries in the neck.70 Cystic infarcts
old man. (C) A slit-like microinfarct (arrows) perpendicular to the pial surface (left),
can also occur as wedge-shaped regions of with a larger triangular region of necrosis at its base near the cortexwhite matter
encephalomalacia in the border zone between junction. (D) Small old linear parenchymal infarct (arrows) filled with macrophages.
two large territories of supply by cerebral
arteries. As discussed above, strategically placed
small infarcts might be just as likely to lead to
dementia as larger infarcts.1,71 A
Lacunar infarcts have historically been
attributed to arteriosclerotic microangiopathy
affecting cerebral parenchymal arteries, a
process sometimes described as lipohyalinosis
and linked etiologically to longstanding hyper-
tension, although that association has recently
been called into question.28,72 Therefore, the
term hypertensive microvascular disease as
a synonym for lipohyalinosis is best avoided.
Lacunar infarcts (Figure 3) have been further
classified into various subtypes on the basis of
their neuropathologic components; the different
types include Ia (small cystic cavity containing B
small blood vessels and a few macrophages),
Ib (incomplete necrosis with perivascular rare-
faction and patchy astrocytic gliosis), and II (an
old microhemorrhage with abundant hemo-
siderin-laden macrophages).72 Hippocampal
injury, often resembling hippocampal scle-
rosis, identified in the medial temporal lobes of Figure 3 Lacunar infarcts. (A) Lacunar infarct
patients with intractable temporal lobe epilepsy (arrowhead) in the left basal ganglia of a patient
(with severe segmental neuron loss in the CA1 with dementia. Larger cystic infarcts (arrows) are
segment of the pyramidal cell layer and frequent seen at the junction of the right basal ganglia with
the adjacent white matter, and in the white matter
preservation of CA2) is now recognized as a
of the right centrum semiovale. (B) Lacunar infarcts
common neuropathologic finding in the CNS in the pons of a different patient (arrows). Lacunar
of elderly patients with dementia, and should infarcts are thought to be one consequence of
probably be considered as part of the spectrum longstanding hypertension, although the strength
of VaD.6,73 of this association has recently been questioned.

OCTOBER 2006 VOL 2 NO 10 SELNES AND VINTERS NATURE CLINICAL PRACTICE NEUROLOGY 543

2006 Nature Publishing Group

 REVIEW
www.nature.com/clinicalpractice/neuro
A B
C D
E amyloid- (A) protein (especially A140,
Figure 4 Cerebral amyloid angiopathy, sporadic, with extensive old
microinfarcts throughout the CNS. The specimens were taken from a
77-year-old patient who was considered clinically to have Alzheimers
disease (AD), and who showed significant AD changes in the hippocampus
and neocortex at autopsy. (A,B) Hematoxylin and eosin-stained sections (at
low and high magnification, respectively) show severe thickening of cerebral
parenchymal arterioles, in which the media has been replaced by eosinophilic
glassy hyaline material (arrows). (C) Amyloid- (A) immunohistochemistry
shows replacement of many arteriolar walls by A-immunoreactive material
(arrows). (D) Multiple remote microinfarcts (similar to those indicated by
arrows) and microhemorrhages were found throughout all areas of the cortex.
(E) Infarcts were highlighted by CD68 and glial fibrillary acidic protein (GFAP)
immunohistochemistry (this panel shows anti-GFAP immunoreactivity).
Cerebral amyloid angiopathy
Cerebral amyloid angiopathy (CAA) describes
an enigmatic microangiopathy that affects
meningeal and cortical arterioles, venules and
capillaries, whereby the normal vessel wall
becomes replaced by fibrillar amyloid.7476 The
most common form of CAA is an age-related
vasculopathy that appears to be confined to the
brain. CAA is strongly associated with AD, such
544 NATURE CLINICAL PRACTICE NEUROLOGY
2006 Nature Publishing Group
that almost all AD patients have some degree of
CAA when this is diligently sought by multiple
sections from autopsy brains.7779 In rare cases,
individuals with prominent CAA lack other
neuropathologic characteristics of AD (senile
plaques and neurofibrillary tangles), and might
not show signs of dementia during their life-
time.80 Severe CAA, in which the medial layer
of many or most cortical arterioles is completely
replaced by amyloid, is also comparatively rare.
Replacement of the smooth muscle cell layer
of the arteriolar media by fibrillar amyloid
renders the vessel susceptible to spontaneous
rupture, which can occur with mild increases
in intravascular pressure or when thrombo-
lytic agents are administered.76 This can result
in nontraumatic lobar cerebral hemorrhage,
although this is an infrequent occurrence
considering the vast numbers of individuals
who have some degree of CAA.74,77 In biopsy
or autopsy brain specimens, antibodies to
as opposed to A142) effectively immuno-
label arteriolar and capillary walls, showing
that vascular amyloid in CAA is similar to the
amyloid in senile plaques.75
A small subset of patients with sporadic or
age-related CAA develop superimposed vascu-
litis. This granulomatous vasculitis, often with
a significant giant cell component, appears to
result from amyloid deposition in arteriolar
walls, rather than being a cause of amyloid
deposition.81 Recent studies have indicated
that vasculitis with CAA might present with a
clinically distinctive syndrome of rapid cogni-
tive decline and seizures, rather than cerebral
hemorrhage.82 There is also growing interest
in the likelihood that patients with severe CAA
might represent an underappreciated variant
of VaD (Figure 4).83 In our California-wide
longitudinal study of individuals at high risk
for VaD, as many as 810% of those exam-
ined at necropsy showed severe CAA, often
with associated cortical microinfarcts, severe
subcortical leukoencephalopathy or both.
It might be predicted that all of these indi-
viduals would also have advanced (Braak
stage VVI) AD, but instead, a subset of them
showed Alzheimerization of the CNS that did
not extend to the isocortical stage. Further
characterization of this interesting patient
group is in progress. In large casecontrol
studies, A CAA has been established as a risk
factor for cerebral ischemic infarcts.84
SELNES AND VINTERS OCTOBER 2006 VOL 2 NO 10

An autosomal dominant syndrome of
dementia and stroke (cerebral hemorrhage),
characterized by extensive, often over-
whelming and widely distributed meningo-
cortical CAA, results from a point mutation in
codon 693 of the gene that codes for amyloid
precursor protein (APP). The resulting
disease, which is observed in circumscribed
regions of the Netherlands, is described as here-
ditary cerebral hemorrhage with amyloidosis,
Dutch-type (HCHWA-D).74,85 Interestingly,
dementia in these patients clearly appears to
be linked to the severity or burden of CAA
rather than to the density of senile plaques
and neurofibrillary tangles.86 This indicates
that sporadic, age-related or AD-associated
CAA, especially when severe, might be a
significantperhaps even definingfactor
in the cognitive decline in this disorder. Other
familial syndromes in which A-immuno-
reactive CAA is a dominant neuropathologic
finding result from mutations in codons
692694 of the APP gene.74,76 Very recently, a
neuropathologically pure form of CAA was
identified in association with an autosomal
dominant APP L705V point mutation within
the A sequence.87
CONCLUSIONS
The field of cerebrovascular disease and
cognitive impairment has made significant
progress over the past decade, including the
discovery of CADASIL, the first genetic form
of a vascular cognitive disorder. Several chal-
lenges lie ahead, however. Population-based
epidemiological studies suggest a continuum
of AD-type pathology and vascular brain
disease, and the relatively arbitrary separation
of AD and vascular dementia may no longer be
the most productive approach. Improvements
in neuroimaging have demonstrated that brain
microvascular disease becomes increasing ly
common with advancing age, but the poor
correlation with degree of cognitive impair-
ment indicates that better measures to capture
the total burden of vascular disease of the
brain are needed. Ultimately, the importance
of better understanding of the pathophysiology
of dementia with cerebrovascular disease
is the potential that some of the risk factors
might be treatable, so that the incidence of
this age-related disorder might eventually
be reduced.
OCTOBER 2006 VOL 2 NO 10 SELNES AND VINTERS
2006 Nature Publishing Group
REVIEW
www.nature.com/clinicalpractice/neuro
KEY POINTS
Several clinical subtypes of vascular cognitive
impairment due to vascular disease are now
recognized, ranging in severity from mild
cognitive impairment to dementia
The risk of dementia after stroke is increased
in older patients and in those with pre-existing
cognitive impairment or cerebrovascular
disease
The cognitive consequences of subcortical
small-vessel disease is variable, suggesting
that the subcortical white matter changes
visualized on conventional MRI might be
incomplete markers of the total burden of
cerebrovascular disease
The cognitive profile of vascular cognitive
impairment is predominantly subcortical, with
prominent psychomotor slowing and executive
deficits, but relatively preserved language and
recognition memory
From a neuropathologic perspective, the
morphologic substrates or correlates of
vascular dementia are extremely heterogeneous
Cerebral microvascular diseases, both
sporadic and genetically determined, including
cerebral autosomal dominant arteriopathy with
subcortical infarcts and leukoencephalopathy
(CADASIL) and cerebral amyloid angiopathy,
are instructive in terms of providing pathologic
models of age-related arteriopathies and their
consequences for brain parenchyma
The brains of aged individuals are likely to
show impaired function as a consequence
of combined Alzheimerization and
cerebrovascular disease
References
1 Chui H (2005) Neuropathology lessons in vascular
dementia. Alzheimer Dis Assoc Disord 19: 4552
2 Fernando MS and Ince PG (2004) Vascular pathologies
and cognition in a population-based cohort of elderly
people. J Neurol Sci 226: 1317
3 Neuropathology Group of the Medical Research
Council Cognitive Function and Ageing Study (2001)
Pathological correlates of late-onset dementia in a
multicentre, community-based population in
England and Wales. Neuropathology Group of the
Medical Research Council Cognitive Function and
Ageing Study (MRC CFAS). Lancet 357:
169175
4 OBrien JT et al. (2003) Vascular cognitive impairment.
Lancet Neurol 2: 8998
5 Rockwood K et al. (2005) Clinical and radiographic
subtypes of vascular cognitive impairment in a clinic-
based cohort study. J Neurol Sci 240: 714
6 Vinters HV et al. (2000) Neuropathologic substrates of
ischemic vascular dementia. J Neuropathol Exp Neurol
59: 931945
NATURE CLINICAL PRACTICE NEUROLOGY 545
REVIEW
www.nature.com/clinicalpractice/neuro
7 Pohjasvaara T et al. (2000) Comparison of different
clinical criteria (DSM-III, ADDTC, ICD-10, NINDS-
AIREN, DSM-IV) for the diagnosis of vascular
dementia. National Institute of Neurological Disorders
and Stroke-Association Internationale pour la
Recherche et lEnseignement en Neurosciences.
Stroke 31: 29522957
8 Bowler JV (2002) The concept of vascular cognitive
impairment. J Neurol Sci 203204: 1115
9 Sachdev P (1999) Vascular cognitive disorder. Int J
Geriatr Psychiatry 14: 402403
10 Roman GC et al. (2004) Vascular cognitive disorder: a
new diagnostic category updating vascular cognitive
impairment and vascular dementia. J Neurol Sci 226:
8187
11 Roman GC et al. (2002) Subcortical ischaemic
vascular dementia. Lancet Neurol 1: 426436
12 Leys D et al. (2005) Poststroke dementia. Lancet
Neurol 4: 752759
13 Linden T et al. (2004) Cognitive impairment and
dementia 20 months after stroke. Neuroepidemiology
23: 4552
14 Desmond DW et al. (2002) Incidence of dementia after
ischemic stroke: results of a longitudinal study. Stroke
33: 22542260
15 Henon H et al. (2001) Poststroke dementia: incidence
and relationship to prestroke cognitive decline.
Neurology 57: 12161222
16 Luchsinger JA et al. (2005) Aggregation of vascular
risk factors and risk of incident Alzheimer disease.
Neurology 65: 545551
17 Barba R et al. (2000) Poststroke dementia: clinical
features and risk factors. Stroke 31: 14941501
18 Pohjasvaara T et al. (2000) How complex interactions
of ischemic brain infarcts, white matter lesions, and
atrophy relate to poststroke dementia. Arch Neurol 57:
12951300
19 Lin JH et al. (2003) Prediction of poststroke dementia.
Neurology 61: 343348
20 Rowan E et al. (2005) Impact of hypertension and
apolipoprotein E4 on poststroke cognition in subjects
>75 years of age. Stroke 36: 18641868
21 Vermeer SE et al. (2003) Silent brain infarcts and the
risk of dementia and cognitive decline. N Engl J Med
348: 12151222
22 Perren F et al. (2005) The syndrome of combined polar
and paramedian thalamic infarction. Arch Neurol 62:
12121216
23 Madureira S et al. (1999) A follow-up study of cognitive
impairment due to inferior capsular genu infarction.
J Neurol 246: 764769
24 Longstreth WT Jr (2005) Brain vascular disease overt
and covert. Stroke 36: 20622063
25 Gold G et al. (1998) Re-evaluating the role of vascular
changes in the differential diagnosis of Alzheimers
disease and vascular dementia. Eur Neurol 40: 121129
26 Ott A et al. (1999) Diabetes mellitus and the risk of
dementia: The Rotterdam Study. Neurology 53:
19371942
27 Langa KM et al. (2004) Mixed dementia: emerging
concepts and therapeutic implications. JAMA 292:
29012908
28 Kalimo H (Ed.; 2005) Pathology and Genetics:
Cerebrovascular Diseases. Basel: ISN Neuropath Press
29 Jellinger KA (2002) The pathology of ischemic-
vascular dementia: an update. J Neurol Sci 203204:
153157
30 Hansen LA and Crain BJ (1995) Making the diagnosis
of mixed and non-Alzheimers dementias. Arch Pathol
Lab Med 119: 10231031
31 Nolan KA et al. (1998) Absence of vascular dementia in
an autopsy series from a dementia clinic. J Am Geriatr
Soc 46: 597604
546 NATURE CLINICAL PRACTICE NEUROLOGY
2006 Nature Publishing Group
32 Longstreth WT Jr et al. (2005) Incidence,
manifestations, and predictors of worsening white
matter on serial cranial magnetic resonance imaging
in the elderly: the Cardiovascular Health Study. Stroke
36: 5661
33 Schmidtke K and Hull M (2005) Cerebral small vessel
disease: how does it progress? J Neurol Sci 229230:
1320
34 Ringelstein EB and Nabavi DG (2005) Cerebral small
vessel diseases: cerebral microangiopathies. Curr
Opin Neurol 18: 179188
35 Goto T et al. (2001) Magnetic resonance imaging
findings and postoperative neurologic dysfunction in
elderly patients undergoing coronary artery bypass
grafting. Ann Thorac Surg 72: 137142
36 Udaka F et al. (2002) White matter lesions and
dementia: MRIpathological correlation. Ann NY Acad
Sci 977: 411415
37 Kuller LH et al. (2005) Determinants of vascular
dementia in the Cardiovascular Health Cognition
Study. Neurology 64: 15481552
38 Prins ND et al. (2005) Cerebral small-vessel disease
and decline in information processing speed, executive
function and memory. Brain 128: 20342041
39 de Groot JC et al. (2000) Cerebral white matter lesions
and cognitive function: the Rotterdam Scan Study. Ann
Neurol 47: 145151
40 Duning T et al. (2005) Excellent cognitive performance
despite massive cerebral white matter changes.
Neuroradiology 47: 749752
41 de Leeuw FE et al. (2004) Alzheimers diseaseone
clinical syndrome, two radiological expressions:
a study on blood pressure. J Neurol Neurosurg
Psychiatry 75: 12701274
42 Mosley TH Jr et al. (2005) Cerebral MRI findings and
cognitive functioning: the Atherosclerosis Risk in
Communities study. Neurology 64: 20562062
43 Roman GC (2004) Brain hypoperfusion: a critical factor
in vascular dementia. Neurol Res 26: 454458
44 Kalaria RN et al. (2001) Multiple substrates of late-
onset dementia: implications for brain protection.
Novartis Found Symp 235: 4960
45 de Leeuw FE et al. (2000) Atrial fibrillation and the
risk of cerebral white matter lesions. Neurology 54:
17951801
46 Markus HS et al. (2000) Reduced cerebral blood flow in
white matter in ischaemic leukoaraiosis demonstrated
using quantitative exogenous contrast based perfusion
MRI. J Neurol Neurosurg Psychiatry 69: 4853
47 Werring DJ et al. (2004) Cognitive dysfunction in
patients with cerebral microbleeds on T2*-weighted
gradient-echo MRI. Brain 127: 22652275
48 Kovari E et al. (2004) Cortical microinfarcts and
demyelination significantly affect cognition in brain
aging. Stroke 35: 410414
49 Vinters HV (2001) Cerebral amyloid angiopathy: a
microvascular link between parenchymal and vascular
dementia? Ann Neurol 49: 691693
50 Chabriat H and Bousser MG (2003) CADASIL. Cerebral
autosomal dominant arteriopathy with subcortical
infarcts and leukoencephalopathy. Adv Neurol 92:
147150
51 Dichgans M (2002) CADASIL: a monogenic condition
causing stroke and subcortical vascular dementia.
Cerebrovasc Dis 13 (Suppl 2): 3741
52 Miao Q et al. (2004) Fibrosis and stenosis of the
long penetrating cerebral arteries: the cause of
the white matter pathology in cerebral autosomal
dominant arteriopathy with subcortical infarcts and
leukoencephalopathy. Brain Pathol 14: 358364
53 Opherk C et al. (2004) Long-term prognosis and
causes of death in CADASIL: a retrospective study in
411 patients. Brain 127: 25332539
SELNES AND VINTERS OCTOBER 2006 VOL 2 NO 10

54 Arango-Lasprilla JC et al. (2005) Cortical and
subcortical diseases: do true neuropsychological
differences exist? Arch Clin Neuropsychol 21: 2940
55 Laukka EJ et al. (2004) Similar patterns of cognitive
deficits in the preclinical phases of vascular dementia
and Alzheimers disease. J Int Neuropsychol Soc 10:
382391
56 Frisoni GB et al. (2002) Mild cognitive impairment with
subcortical vascular features: clinical characteristics
and outcome. J Neurol 249: 14231432
57 Graham NL et al. (2004) Distinctive cognitive profiles
in Alzheimers disease and subcortical vascular
dementia. J Neurol Neurosurg Psychiatry 75: 6171
58 Tierney MC et al. (2001) Recognition memory and
verbal fluency differentiate probable Alzheimer disease
from subcortical ischemic vascular dementia. Arch
Neurol 58: 16541659
59 Garrett KD et al. (2004) The neuropsychological profile
of vascular cognitive impairmentno dementia:
comparisons to patients at risk for cerebrovascular
disease and vascular dementia. Arch Clin
Neuropsychol 19: 745757
60 Fahlander K et al. (2002) Cognitive functioning in
Alzheimers disease and vascular dementia: further
evidence for similar patterns of deficits. J Clin Exp
Neuropsychol 24: 734744
61 Looi JC and Sachdev PS (1999) Differentiation of
vascular dementia from AD on neuropsychological
tests. Neurology 53: 670678
62 Traykov L et al. (2002) Neuropsychological deficit in
early subcortical vascular dementia: comparison to
Alzheimers disease. Dement Geriatr Cogn Disord 14:
2632
63 Peters N et al. (2005) The pattern of cognitive
performance in CADASIL: a monogenic condition
leading to subcortical ischemic vascular dementia. Am
J Psychiatry 162: 20782085
64 Buffon F et al. (2006) Cognitive profile in CADASIL.
J Neurol Neurosurg Psychiatry 77: 175180
65 OSullivan M et al. (2004) Diffusion tensor imaging of
thalamus correlates with cognition in CADASIL without
dementia. Neurology 62: 702707
66 Reed BR et al. (2004) Clinical and neuropsychological
features in autopsy-defined vascular dementia. Clin
Neuropsychol 18: 6374
67 Vinters HV et al. (1998) Diagnostic Neuropathology,
149. New York: Marcel Dekker Inc.
68 Medana IM and Esiri MM (2003) Axonal damage: a key
predictor of outcome in human CNS diseases. Brain
126: 515530
69 Brown WR et al. (2000) Apoptosis in leukoaraiosis.
AJNR Am J Neuroradiol 21: 7982
70 Hulette C et al. (1997) Clinical-neuropathologic
findings in multi-infarct dementia: a report of six
autopsied cases. Neurology 48: 668672
OCTOBER 2006 VOL 2 NO 10 SELNES AND VINTERS
2006 Nature Publishing Group
REVIEW
www.nature.com/clinicalpractice/neuro
71 Del ST et al. (1990) Vascular dementia: a Acknowledgments
clinicopathological study. J Neurol Sci 96: 117 Work in H Vinters laboratory
72 Lammie GA (2002) Hypertensive cerebral small vessel is supported by PHS grants
disease and stroke. Brain Pathol 12: 358370 P50 AG 16570 and P01
73 Dickson DW et al. (1994) Hippocampal sclerosis: a AG 12435. The research of
common pathological feature of dementia in very old O Selnes is supported by
(> or = 80 years of age) humans. Acta Neuropathol PHS grant 35610 (NINCDS).
(Berl) 88: 212221 Technical assistance
74 Verbeek MM et al. (2000) Cerebral Amyloid Angiopathy provided by Justine
in Alzheimers Disease and Related Disorders. Pomakian and Negar
Dordrecht: Kluwer Academic Publishers Khanlou. We thank Pamela
75 Revesz T et al. (2002) Sporadic and familial cerebral Talalay for her editorial
amyloid angiopathies. Brain Pathol 12: 343357 assistance.
76 Rensink AA et al. (2003) Pathogenesis of cerebral
amyloid angiopathy. Brain Res Brain Res Rev 43: Competing interests
207223 The authors declared
77 Vinters HV (1987) Cerebral amyloid angiopathy: they have no competing
a critical review. Stroke 18: 311324 interests.
78 Vinters HV et al. (1996) Brain parenchymal and
microvascular amyloid in Alzheimers disease. Brain
Pathol 6: 179195
79 Greenberg SM et al. (1993) The clinical spectrum of
cerebral amyloid angiopathy: presentations without
lobar hemorrhage. Neurology 43: 20732079
80 Vinters HV (1992) Cerebral amyloid angiopathy and
Alzheimers disease: two entities or one? J Neurol Sci
112: 13
81 Anders KH et al. (1997) Giant cell arteritis in
association with cerebral amyloid angiopathy:
immunohistochemical and molecular studies. Hum
Pathol 28: 12371246
82 Eng JA et al. (2004) Clinical manifestations of cerebral
amyloid angiopathy-related inflammation. Ann Neurol
55: 250256
83 Haglund M et al. (2004) Severe cerebral amyloid
angiopathy characterizes an underestimated variant
of vascular dementia. Dement Geriatr Cogn Disord 18:
132137
84 Cadavid D et al. (2000) Cerebral beta amyloid
angiopathy is a risk factor for cerebral ischemic
infarction: a case control study in human brain
biopsies. J Neuropathol Exp Neurol 59: 768773
85 Vinters HV et al. (1998) Secondary microvascular
degeneration in amyloid angiopathy of patients with
hereditary cerebral hemorrhage with amyloidosis,
Dutch type (HCHWA-D). Acta Neuropathol (Berl) 95:
235244
86 Natte R et al. (2001) Dementia in hereditary cerebral
hemorrhage with amyloidosis-Dutch type is
associated with cerebral amyloid angiopathy but is
independent of plaques and neurofibrillary tangles.
Ann Neurol 50: 765772
87 Obici L et al. (2005) A novel APP mutation exclusively
associated with cerebral amyloid angiopathy. Ann
Neurol 58: 639644
NATURE CLINICAL PRACTICE NEUROLOGY 547