Hippocampal Neuronal Atrophy and Cognitive Function in Delayed Poststroke and Aging-Related Dementias
We have previously shown delayed poststroke dementia in elderly (≥75 years old) stroke survivors is associated with medial temporal lobe atrophy; however, the basis of the structural and functional changes is unknown. Using 3-dimensional stereological methods, we quantified hippocampal pyramidal neuronal volumes and densities in a total of 95 postmortem samples from demented and nondemented poststroke survivors within our prospective Cognitive Function after Stroke study and subjects pathologically diagnosed with vascular dementia, Alzheimer disease, and mixed Alzheimer disease and vascular dementia syndrome. Hippocampal CA1 but not CA2 subfield neuron density was affected in poststroke, Alzheimer disease, vascular dementia, and mixed dementia groups relative to control subjects (P<0.05). Neuronal volume was reduced in the poststroke dementia relative to poststroke nondemented group in both CA1 and CA2, although there were no apparent differences in neuronal density. Poststroke nondemented neuronal volumes were similar to control subjects but greater than in all dementias (P<0.05). Neuronal volumes positively correlated with global cognitive function and memory function in both CA1 and CA2 in poststroke subjects (P<0.01). Degrees of neuronal atrophy and loss were similar in the poststroke dementia and vascular dementia groups. However, in the entorhinal cortex layer V, neuronal volumes were only impaired in the mixed and Alzheimer disease groups (P<0.05). Our results suggest hippocampal neuronal atrophy is an important substrate for dementia in both cerebrovascular and neurodegenerative disease.
Hippocampal Neuronal Atrophy and Cognitive Function in
Delayed Poststroke and Aging-Related Dementias
Elizabeth Gemmell, MRes; Helen Bosomworth, MRes; Louise Allan, PhD; Roslyn Hall, BSc;
Ahmad Khundakar, PhD; Arthur E. Oakley, BSc; Vincent Deramecourt, MD;
Tuomo M. Polvikoski, MD; John T. O’Brien, DM; Raj N. Kalaria, FRCPath
Background and Purpose—We have previously shown delayed poststroke dementia in elderly (ⱖ75 years old) stroke
survivors is associated with medial temporal lobe atrophy; however, the basis of the structural and functional changes
Methods—Using 3-dimensional stereological methods, we quantified hippocampal pyramidal neuronal volumes and
densities in a total of 95 postmortem samples from demented and nondemented poststroke survivors within our
prospective Cognitive Function after Stroke study and subjects pathologically diagnosed with vascular dementia,
Alzheimer disease, and mixed Alzheimer disease and vascular dementia syndrome.
Results—Hippocampal CA1 but not CA2 subfield neuron density was affected in poststroke, Alzheimer disease, vascular
dementia, and mixed dementia groups relative to control subjects (P⬍0.05). Neuronal volume was reduced in the
poststroke dementia relative to poststroke nondemented group in both CA1 and CA2, although there were no apparent
differences in neuronal density. Poststroke nondemented neuronal volumes were similar to control subjects but greater
than in all dementias (P⬍0.05). Neuronal volumes positively correlated with global cognitive function and memory
function in both CA1 and CA2 in poststroke subjects (P⬍0.01). Degrees of neuronal atrophy and loss were similar in
the poststroke dementia and vascular dementia groups. However, in the entorhinal cortex layer V, neuronal volumes
were only impaired in the mixed and Alzheimer disease groups (P⬍0.05).
Conclusions—Our results suggest hippocampal neuronal atrophy is an important substrate for dementia in both
cerebrovascular and neurodegenerative disease. (Stroke. 2012;43:808-814.)
Key Words: Alzheimer disease 䡲hippocampus 䡲poststroke dementia 䡲stroke 䡲vascular dementia
Patients with stroke who do not develop dementia as a
direct result of a stroke have a 9-fold increased risk of
developing delayed poststroke dementia (PSD),
affect up to 50% of all stroke survivors and is associated with
poor long-term survival.
However, the pathological pro-
cesses that increase vulnerability of nondemented stroke
survivors to cognitive decline are unknown. We have previ-
ously shown that medial temporal lobe atrophy (MTA) on
MRI was a predictor of delayed PSD in elderly patients with
stroke who were not cognitively impaired 3 months post-
Given that hippocampal atrophy is considered a
sensitive marker of Alzheimer disease (AD),
that MTA reflects the presence of asymptomatic Alzheimer
pathology uncovered or exacerbated by ischemic injury.
However, there is growing evidence for MTA and hippocam-
pal degeneration associated with cerebrovascular disease in
ischemic vascular dementia (VaD)
and dementia caused
by sporadic and familial small vessel disease,
postmortem verification of neurodegenerative pathology was
not available in most of these studies.
The basis of MTA and associated functional impairment in
delayed PSD is unclear. In AD, the cause is generally thought
to be hippocampal neurodegeneration, particularly in the CA1
This may also be true in PSD because CA1
neurons are vulnerable to ischemia and hypoperfusion,
and human and experimental studies suggest there is signif-
icant CA1 neuron loss after ischemic stroke.
rodegeneration in stroke survivors who developed delayed
PSD may therefore account for more severe cognitive impair-
ment compared with those who maintained normal cognitive
function. However, loss of neurons is not the only factor that
contributes to brain atrophy and functional decline. This has
been demonstrated in the hippocampus, where although
atrophy is strongly associated with memory impairment in
Received August 23, 2011; final revision received October 12, 2011; accepted October 26, 2011.
From the Centre for Brain Ageing and Vitality, Institute for Ageing and Health, Newcastle University, Campus for Ageing & Vitality, Newcastle upon
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/
Correspondence to Raj N. Kalaria, FRCP, Institute for Ageing and Health, Campus for Ageing & Vitality, Newcastle upon Tyne, NE4 5PL, UK. E-mail
© 2011 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org DOI: 10.1161/STROKEAHA.111.636498
AD and normal aging,
there is limited neuron loss with
and neurodegeneration in AD makes only a weak
contribution to tissue atrophy.
reports of CA1 neuron loss in VaD suggest other mechanisms
Volumetric or atrophic changes in neuro-
nal morphometry have also been suggested as a major
contributory factor to gross structural and functional changes
This study therefore investigated the impact
of cellular pathology on cognitive dysfunction in cerebrovas-
cular and neurodegenerative diseases associated with MTA
and dementia. Hippocampal pyramidal neuron density and
soma volume were examined in relation to cognitive function
in prospectively assessed stroke survivors who had developed
delayed PSD or remained nondemented.
Subjects and Methods
Subject Selection and Clinical Diagnosis
The demographic details of the different subjects and pathological
findings are presented in Table 1. Hippocampal tissue was analyzed
from 36 poststroke subjects from the prospective Cognitive Function
After Stroke (CogFAST) study as described previously.
patients ⱖ75 years old were selected if they were not demented 3
months poststroke and did not exhibit disabilities that would prevent
them from completing cognitive tests. They received annual clinical
assessments and a neuropsychological test battery from baseline
including the Cognitive Drug Research battery, the Mini-Mental
State Examination, and the Cambridge Assessment of Mental Dis-
orders in the Elderly, which generated subscores for various cogni-
tive domains, including memory and executive function.
Subjects were diagnosed as demented if they met Diagnostic and
Statistical Manual of Mental Disorders, Third Edition Revised
criteria for dementia. Control subjects ⬎75 years old were only
selected if they demonstrated no evidence of cognitive impairment;
however, they were not psychologically tested. Ethical approval was
granted by local research ethics committees for this study (Newcastle
on Tyne Hospitals Trust, UK) and permission for postmortem
research using brain tissue was granted for this project.
Final diagnoses of demented subjects were assigned based on
established neuropathological diagnostic criteria.
ylin and eosin staining was used for assessment of structural integrity
and infarcts, Nissl and Luxol fast blue staining for cellular pattern
and myelin loss, Bielschowsky silver impregnation for Consortium
to Establish a Registry for Alzheimer’s Disease (CERAD) rating of
neuritic plaques, and tau immunohistochemistry for Braak staging of
neurofibrillary tangles. A diagnosis of VaD was made when there
were multiple or cystic infarcts, lacunae, microinfarcts, and small
vessel disease and Braak stage ⬍III.
A diagnosis of AD was made
when there was evidence of significant Alzheimer-type pathology
(Braak Stage V–VI and moderate to severe CERAD score) and
absence of significant vascular pathology. Subjects were diagnosed
as “mixed” when there was evidence of concomitant VaD with AD,
Lewy body pathology, or tauopathy. We found that 70% of delayed
PSD cases met pathological criteria for VaD at autopsy with
negligible neurodegenerative pathology.
Global vascular pathology
was assessed (V.D., R.K.) and calculated as the sum of ratings of
vascular pathology in the hippocampus, frontal lobe, temporal lobe,
and basal ganglia to generate a score /20 (Deramecourt et al,
To investigate the effects of different disease processes, we also
analyzed neurons in elderly control subjects and VaD, AD, and
dementia with mixed AD and VaD pathology cases. Brain tissues
were retrieved from the Newcastle Brain Tissue Resource (New-
castle, UK) except 4 control cases, which were obtained from the
Medical Research Council London Brain Bank for Neurodegenera-
tive Diseases (Institute of Psychiatry, London, UK). Brain tissue was
cut from predefined, paraffin-embedded blocks of the hippocampus
according to the Newcastle Brain Map.
The coronal samples were
taken between the level of the pregeniculate nucleus and the pulvinar
at which the emergence of the ventricle is visible.
Table 1. Demographic Details of Groups Analyzed*
Control PSND PSD VaD Mixed AD
Maximum no. of subjects
14 22 16 13 16 14
Mean (range) 81 (72–92) 84 (78 –94) 88 (80 –98) 86 (71–97) 84 (72–94) 84 (70 –91)
Mean (range) NPD 2.6 (1–5) 2.6 (0– 4) 2.1 (1–4) 5.1 (4– 6) 5.2 (4– 6)
Mean (range) NPD 1.5 (0 –2) 1 (0 –3) 1.3 (0 –2) 2.6 (2–3) 3 (3)
Mean (range) NPD 12.6 (7–16) 11.9 (8 –17) 13 (12–14) 3 (6–14) N/A
Mean (⫾2 SE) 25 (7) 46 (12) 48 (13) 43 (10) 36 (11) 56 (18)
Fixation duration, wk
Mean (⫾2 SE) 15 (4) 11 (3) 8 (2) 12 (6) 17 (5) 10 (3)
Mean (⫾2 SE) 24.1 (1.6) 26.3 (0.3) 27 (0.3) 27.1 (0.6) 26.2 (0.7) 26.2 (0.9)
PSND indicates nondemented poststroke subjects; PSD, delayed poststroke dementia; VaD, vascular dementia, “mixed” Alzheimer,
and vascular dementia; AD, Alzheimer disease; CERAD, Consortium to Establish a Registry for Alzheimer’s Disease; PMD, postmortem
delay; NPD, no pathological diagnoses; ECV, entorhinal cortex Layer V.
*Maximum number in control group ECV n⫽8, PSND CA2 n⫽20, and ECV n⫽20; PSD CA2 n⫽21 and ECV n⫽11; VaD CA2 n⫽12
and ECV n⫽11; mixed CA2 n⫽12 and ECV n⫽11; AD CA2 n⫽13 and ECV n⫽12. Total no. in CA1 n⫽95, in CA2 n⫽82 and ECV
Gemmell et al Neuronal Atrophy and Poststroke Dementia 809
neuroimaging did not detect differences in the degree of MTA in the
left versus right hemisphere; therefore, sections were taken from
either the left or right hippocampus in each case provided there were
no apparent gross lesions. At least 3 30-
m sections per case were
cut using a Shandon FinessE⫹rotary microtome. Sections were
stained using the Nissl method and checked for quality, staining
consistency, and penetration. Slides were coded so the investigators
(E.G., H.B.) were blind to disease group. To minimize differential
tissue effects from processing and staining, all cases were collected,
treated, and analyzed in a standardized manner, allowing accurate
and valid comparisons to be made.
Pyramidal neurons were analyzed in hippocampal areas CA1, CA2,
and entorhinal cortex Layer V (ECV), except 13 cases in which CA2
and 22 cases where ECV was not present or suitable for analysis
(Table 1). Sections were imaged using a Zeiss Axioplan Photomi-
croscope with a Pixelink PL-B623CF color digital camera linked to
a computer. The reference area was defined by the investigator with
a⫻2.5 objective lens using stereological analysis software (Visiop-
harm Integration System, Hørsholm, Denmark). At this magnifica-
tion, CA1 and CA2 were easily distinguishable because CA1 was
broader with lower neuron density than CA2. The entorhinal cortex
was defined according to Insausti et al.
Layer V was identified as
a band of darkly stained large- to medium-sized pyramidal neurons
superficially bordered by cell-sparse Layer IV as described by Canto
A motorized stage (Prior ProScan II; Prior Scientific Instruments
Ltd, Cambridge, UK) with an accuracy of 1
m was used to take up
a uniform random sampling procedure within the reference area to
select approximately 33 frames from each of the 3 sections per case.
Frames were viewed at ⫻100 magnification using an oil immersion
objective with a numeric aperture of 1.25. The optical disector
method was used for estimation of neuron density.
Based on x–y
axis length, each disector frame had an area of 2548.66
Heidenhain z-axis microcator (Heidenhain GB Ltd, London, UK),
accurate to 0.5
m, was used to precisely measure disector depth and
tissue section thickness, which was recorded every 10 frames (mean,
m). Each disector probe had a z-axis depth of 18
excluding a guard volume ⱖ4
m. Pyramidal neurons were identi-
fied using established criteria, that is, characteristic triangular soma,
Nissl-stained cytoplasm, and darkly stained single nucleolus.
pilot study was performed to ensure that the neurons were sampled
in sufficient numbers to provide precise estimates based on the mean
coefficient of error. Approximately 150 neurons were analyzed per
subfield per case, ensuring that the average sampling error reached a
The soma volume of each sampled neuron was
measured using an independent uniform random orientated nucleator
probe (Figure 1).
Statistical analyses were carried out using SPSS Version 17.0.
Significance was set at P⬍0.05. The Shapiro-Wilk test was used to
check for normal distribution and data nonnormally distributed were
analyzed using nonparametric tests. Group means were compared
using the Kruskal-Wallis test. Pairwise analyses were performed
using the Mann-Whitney UTest. The Wilcoxon signed rank test
was used for comparison of related means. Correlations were assessed
using Spearman rank correlation. Pearson
test or Fisher exact tests
were used to determine associations between categorical data.
There were no differences in the mean age or the distribution
of males and females between groups. Poststroke cases were
divided into 2 groups based on cognitive status at their final
assessment (mean 7.6 months before death) to allow compar-
ison between nondemented stroke survivors (PSND) and
those who developed delayed PSD (Tables 1 and 2). We
found no apparent differences in the Braak stage, CERAD
score, vascular pathology scores, or time from stroke to death
between PSND and PSD groups (Table 1). There was no
association between Braak stage ⬎2 and delayed PSD.
Sixty-three percent (n⫽10) of delayed PSD cases met path-
ological criteria for a final diagnosis of VaD; the remainder
were mixed AD and VaD. There were no associations
between lesion location and delayed PSD (Fisher exact test
P⫽0.743). The coefficient of error values for neuronal
densities were CA1 P⫽0.06, CA2 P⫽0.002, ECV P⫽0.02,
and for neuronal volumes were CA1 P⫽0.04, CA2 P⫽0.02,
ECV P⫽0.05 demonstrating a high level of precision. Period
of postmortem delay or length of fixation (up to 40 weeks)
was not found to influence neuronal density or volume results
among the various groups (P⬎0.05).
As shown in Figure 2, neuronal density was different between
all regions CA2⬎ECV⬎CA1 (P⬍0.001) and positively cor-
related between CA1 and CA2 and CA1 and ECV (r⫽0.311,
Figure 1. The nucleator method for neuronal volume. Six ran-
domly oriented rays originating from the nucleolus were marked
where they crossed the border of the neuronal soma, and soma
volume was calculated using the nucleator.
P indicates pyram-
idal cell; N, nonpyramidal cell; G, glial cell. Red and green lines
delineate the disector frame. Bar⫽10
Table 2. Clinical Findings in Poststroke Subjects
Time from baseline- death (mo)
Mean (⫾2 SE) 63.5 (22) 64.4 (14)
Total CAMCOG score (/100)
Mean (range) 88.6 (76–99) 64 (24–80)
Memory subscore (/27)
Mean (⫾2 SE) 17 (1.6) 10 (2)
Executive function subscore (/28)
Mean (⫾2 SE) 22 (0.8) 16 (2.3)
Hemisphere with visible lesion on CT
(right, left, both, none) (5, 2, 5, 6) (3, 6, 2, 3)
PSND indicates nondemented poststroke subjects; PSD, delayed poststroke
dementia; CAMCOG, Cambridge Assessment of Mental Disorders in the Elderly.
810 Stroke March 2012
P⫽0.005 and r⫽0.247, P⫽0.037, respectively). CA1 neuro-
nal density was different between the groups (P⬍0.001).
Compared with control subjects, CA1 neuronal density was
reduced in the PS group (P⫽0.025), PSND (P⫽0.027), VaD
(P⫽0.012), mixed (P⬍0.001), and AD groups (P⫽0.001).
In CA1, the mixed dementia and AD groups had lower
neuronal density than the PSND group (P⬍0.001 and
P⫽0.015, respectively). Surprisingly, there was no difference
in CA1 neuronal density between the PSD and PSND groups
(P⫽0.643). CA2 and ECV neuronal densities were not
different between groups (P⫽0.562 and P⫽0.303, respec-
tively); however, there were trends for lower ECV neuronal
density in the mixed group compared with control subjects
and PSND groups (P⫽0.083 and P⫽0.08, respectively).
We found neuronal volume was different across all areas
(CA2⬎CA1⬎ECV; (P⬍0.001; Figure 3). Providing consis-
tency in our results, neuronal volume was correlated in all 3
areas (CA1 versus CA2 r⫽0.406, P⬍0.001; CA1 versus ECV
r⫽0.231, P⫽0.05; CA2 versus ECV r⫽0.311 P⫽0.012).
CA1 neuronal volume was different among all groups
(P⫽0.015). Compared with control subjects, CA1 neuronal
volume was reduced in VaD (P⫽0.047), mixed dementia
(P⫽0.039), and AD (P⫽0.037). Compared with the PSND
group, CA1 neuronal volume was reduced in all dementia
groups (PSD P⫽0.028, VaD P⫽0.026, mixed P⫽0.009, and
AD P⫽0.01). However, there was no difference in neuronal
volume between the PSND group and control subjects
CA2 neuronal volume was different among all groups
(P⫽0.002). Compared with control subjects, CA2 neuronal
volume was reduced in all dementia groups (PSD P⫽0.014,
VaD P⫽0.021, mixed P⫽0.003, and AD P⫽0.016). Com-
pared with the PSND group, CA2 neuron volume was
reduced in all dementia groups (PSD P⫽0.009, mixed
P⫽0.003, and AD P⫽0.019), and there was a trend to
significance with the VaD group (P⫽0.08). We did not find
any differences between the PSND and control subjects
(P⫽0.959). A negative correlation was found between CA2
neuronal volume and age (r⫽⫺0.246, P⫽0.031).
ECV neuronal volumes were different among all groups
(P⬍0.001). ECV neuron volumes were reduced compared
with control subjects and PSND in mixed dementia (P⫽0.001
and P⬍0.001, respectively) and AD (P⫽0.008 and P⫽0.019,
respectively). There were no differences found among the PS,
VaD, and control groups. We did not find differences in
neuronal volume or density between the PSD cases that were
classified as “VaD” or “mixed” (data not shown). There were
no differences in neuronal volume or density between male
and female subjects.
Poststroke Neuronal Volume and
Final total Cambridge Assessment of Mental Disorders in the
Elderly scores were positively correlated with neuronal vol-
ume in CA1 (r⫽0.399, P⫽0.01) and CA2 (r⫽0.445,
P⫽0.007; Table 2; Figure 4). CA2 neuronal volume was
positively correlated with total memory (r⫽0.481, P⫽0.006),
recent memory (r⫽0.692, P⬍0.001), remote memory
(r⫽0.428, P⫽0.016) and attention Cambridge Assessment of
Figure 2. Neuronal densities in hippocampal subregions CA1, CA2, and entorhinal cortex Layer V (ECV). All PS indicates all poststroke
subjects; PSND, nondemented poststroke subjects; PSD, delayed poststroke dementia; VaD, vascular dementia, “mixed” mixed Alzhei-
mer, and vascular dementia; AD, Alzheimer disease. Means⫾2 SE. Asterisks indicate signiﬁcantly different to controls (black) or PSND
(gray; P⬍0.05). Dots indicate trend to signiﬁcance (P⬍0.01).
Figure 3. Neuronal volumes in hippocampal
subregions CA1 , CA2, and entorhinal cortex
Layer V (ECV). All PS indicates all poststroke
subjects; PSND, nondemented poststroke
subjects; PSD, delayed poststroke dementia;
VaD, vascular dementia, “mixed” mixed Alz-
heimer, and vascular dementia; AD, Alzheimer
disease. Means⫾2 SE. Asterisks indicate sig-
niﬁcantly different to controls (black) or PSND
(gray; P⬍0.05). Dots indicate trend to signiﬁ-
Gemmell et al Neuronal Atrophy and Poststroke Dementia 811
Mental Disorders in the Elderly subscores (r⫽0.379,
P⫽0.035). There were trends to significant correlations
between CA2 neuronal volume and learning (r⫽0.352,
P⫽0.052). ECV neuronal volume was correlated with remote
memory subscores (r⫽0.372, P⫽0.043). We did not find
correlations between neuronal density and cognitive function.
CA1 neuron density was negatively correlated with Braak
stage and CERAD score across all groups (r⫽⫺0.379,
P⫽0.002 and r⫽⫺0.0392, P⫽0.001) even where Braak
staging was not diagnostic for AD.
CA2 and ECV neuronal
volume were negatively correlated with Braak stage
(r⫽⫺0.392, P⫽0.003 and r⫽⫺0.395, P⫽0.004) and
CERAD score (r⫽⫺0.261, P⫽0.059 and r⫽⫺0.419,
P⫽0.002). There were no correlations between vascular
pathological burden and neuronal volumes or between neu-
rodegenerative or vascular pathology and neuronal density.
Our results provide novel evidence that reduced neuronal
volume or “neuronal atrophy” is associated with cognitive
impairment in delayed PSD and other aging-related demen-
tias with both vascular etiology and AD. Pyramidal neuronal
volumes in the hippocampal subfields CA1 and CA2 were
10% to 20% smaller in delayed PSD, VaD, mixed, and AD
groups compared with cognitively normal elderly subjects (online-
only Supplementary Table, http://stroke.ahajournals.org). Interest-
ingly, the PSND group had neuronal volumes similar to
control subjects, whereas neuronal volumes in the delayed
PSD group were 20% smaller than PSND. Because the PS
groups had comparable vascular and negligible Alzheimer-
type lesion burden, neuronal atrophy appears as a selective
pathological feature discriminating nondemented and de-
mented elderly stroke survivors. This suggests that reduced
neuronal volume reflects mechanistic change occurring in
some PS survivors accelerating cognitive decline. This was
supported by the positive correlation between neuronal vol-
umes and cognitive test scores in stroke survivors. Further-
more, because neuronal volumes were reduced in VaD, AD,
and mixed dementia, our results suggest that mechanisms
causing neuronal atrophy are associated with cognitive de-
cline in both cerebrovascular disease and neurodegenerative
In agreement with the well-documented differential vul-
nerability of the entorhinal cortex to AD pathology,
nal volumes in Layer V of the entorhinal cortex were
negatively correlated with Braak stage and were reduced in
AD and mixed dementia but not affected in PSD and VaD.
The finding that neuronal volume and density changes in the
delayed PSD group were similar to those in VaD supported
clinicopathological findings that delayed PSD is associated
with vascular-type dementia driven by vascular rather than
Alzheimer or other neurodegenerative pathology.
We found that CA2 neuronal volumes were related to
various memory Cambridge Assessment of Mental Disorders
in the Elderly subscores, which supports a recent study
implicating CA2 neurons in hippocampal function and mem-
Our observations were also consistent with
an MRI study, which found that CA2 atrophy was more
strongly associated with mild cognitive impairment and
progression to dementia than CA1 atrophy in AD.
atrophy appears to be a potential mechanism relating CA2
atrophy to cognitive impairment in the absence of extensive
CA2 neuron loss.
Soma size is thought to reflect the amount of cellular
machinery needed to support the axodendritic arbor.
may therefore conclude that a reduction in neuronal soma
volume in patients with dementia could be related to a loss of
or reduced complexity of dendrites and/or axons.
mechanisms that cause volumetric changes in neuronal mor-
phometry could contribute to disruption of hippocampal
circuitry and cognitive impairment in neurodegenerative and
Our results extend previous studies that suggested neuronal
volume is decreased in AD.
However, our findings differ
from a previous study by Zarow et al
that found no
significant change in hippocampal neuronal volume in AD
and ischemic VaD compared with control subjects. Because
they used a 3-dimensional method similar to ours, we
reasoned that the small numbers of subjects analyzed per
group may have limited their detection of statistically signif-
Appropriate human material is universally difficult to
obtain. However, we were fortunate to have acquired suffi-
cient material of suitable quality to allow accurate analyses
and reliable statistical results. During this study, tissue
sections were taken from predefined paraffin-embedded
blocks with a fixed starting point. This limitation therefore
precluded the strict sampling procedure necessary to conduct
Cavalieri calculations on brain region volume as well as the
Figure 4. Poststroke cognitive function
correlated with CA2 neuronal volumes. A,
CA2 neuronal volumes versus total Cam-
bridge Assessment of Mental Disorders in
the Elderly (CAMCOG) scores (r⫽0.445,
P⫽0.007). B, CA2 neuronal volumes ver-
sus CAMCOG memory subscores
(r⫽0.481, P⫽0.006). x indicates post-
stroke demented; o, poststroke
812 Stroke March 2012
subsequent estimation of total cell number within the struc-
ture using the desirable “fractionator” method.
density measures are based on the relationship between the
numerator (the cell counts) and denominator (extracellular
matrix), the effect of tissue processing cannot be ruled out.
However, all sections were processed and handled in a
standardized manner, allowing valid comparisons to be made
as any introduced artifact would be common to all sections.
Our results were consistent with the literature describing
significant loss of CA1 neurons in VaD and AD.
surprised to find significantly decreased neuronal density in
the CA1 of the PSND group compared with control subjects
and no significant difference in CA1 neuron density between
PS groups because we expected greater neuron loss to explain
progression to PSD. This suggests an important role for
mechanisms causing neuronal atrophy in cognition even after
neuron loss. However, because we have shown that delayed
PSD patients have greater MTA than nondemented stroke
survivors on MRI,
the PSD neuronal density measurements
may be affected by tissue atrophy bringing surviving neurons
closer together. Because the neurodegenerative dementias we
studied are associated with MTA, the true level of neuron loss
is likely to be greater than our results suggested and closer to
values reported in previous studies.
The greatest reduction
in CA1 neuron density was found in the mixed AD and VaD
groups, which supports clinicopathological studies demon-
strating that coexistence of vascular and Alzheimer-type
pathology has an additive effect on cognition.
Our results provide novel evidence of hippocampal neuronal
atrophy in delayed PSD, VaD, mixed AD and VaD, and AD.
ECV neuronal volumes were only reduced in the mixed and
AD groups, which was consistent with the known vulnera-
bility of the ECV to AD pathology. Hippocampal CA1 and
CA2 neuronal volumes were significantly reduced in delayed
PSD subjects compared with nondemented stroke survivors
and were related to global cognitive function as well as
learning and memory function. However, there were no
differences in hippocampal neuronal density between nonde-
mented and demented stroke survivors. Our findings suggest
that even after significant neurodegeneration, therapeutic
strategies to maintain or restore functional morphology in
surviving neurons could prevent further cognitive decline in
PS and aging-related dementias.
We are grateful to the patients, families, and clinical house staff for
their cooperation in the investigation of this study. We thank
Michelle Widdrington, Carein Todd, Jean Scott, Deborah Lett, and
Anne Nicholson for assistance in managing and screening the cohort.
Sources of Funding
Our work is supported by grants from the Newcastle Centre for Brain
Ageing and Vitality (Lifelong Health and Wellbeing programme
funded by Research Councils United Kingdom), Medical Research
Council (MRC) of the United Kingdom (G0500247), and Alzhei-
mer’s Research United Kingdom (ARUK). Tissue for this study was
collected by the Newcastle Brain Tissue Resource, which is funded
in part by a grant from the MRC (G0400074), by the Newcastle
National Institute of Health Research Biomedical Research Centre in
Ageing and Age Related Diseases award to the Newcastle upon Tyne
Hospitals National Health Service Foundation Trust, and by a grant
from the Alzheimer’s Society and ARUK as part of the Brains for
Dementia Research Project.
1. Kokmen E, Whisnant JP, O’Fallon WM, Chu CP, Beard CM. Dementia
after ischemic stroke: a population-based study in Rochester, Minnesota
(1960 –1984). Neurology. 1996;46:154–159.
2. Melkas S, Oksala NK, Jokinen H, Pohjasvaara T, Vataja R, Oksala A, et
al. Poststroke dementia predicts poor survival in long-term follow-up:
influence of prestroke cognitive decline and previous stroke. J Neurol
Neurosurg Psychiatry. 2009;80:865–870.
3. Pendlebury ST, Rothwell PM. Prevalence, incidence, and factors asso-
ciated with pre-stroke and post-stroke dementia: a systematic review and
meta-analysis. Lancet Neurol. 2009;8:1006–1018.
4. Firbank MJ, Burton EJ, Barber R, Stephens S, Kenny RA, Ballard C, et
al. Medial temporal atrophy rather than white matter hyperintensities
predict cognitive decline in stroke survivors. Neurobiol Aging. 2007;28:
5. Frisoni GB, Fox NC, Jack CR Jr, Scheltens P, Thompson PM. The
clinical use of structural MRI in Alzheimer disease. Nat Rev Neurol.
6. Barber R, Gholkar A, Scheltens P, Ballard C, McKeith IG, O’Brien JT.
Medial temporal lobe atrophy on MRI in dementia with Lewy bodies.
7. Bastos-Leite AJ, van der Flier WM, van Straaten EC, Staekenborg SS,
Scheltens P, Barkhof F. The contribution of medial temporal lobe atrophy
and vascular pathology to cognitive impairment in vascular dementia.
8. Scher AI, Xu Y, Korf ES, Hartley SW, Witter MP, Scheltens P, et al.
Hippocampal morphometry in population-based incident Alzheimer’s
disease and vascular dementia: the HAAS. J Neurol Neurosurg Psy-
9. Kril JJ, Patel S, Harding AJ, Halliday GM. Patients with vascular
dementia due to microvascular pathology have significant hippocampal
neuronal loss. J Neurol Neurosurg Psychiatry. 2002;72:747–751.
10. Barber R, Ballard C, McKeith IG, Gholkar A, O’Brien JT. MRI volu-
metric study of dementia with Lewy bodies: a comparison with AD and
vascular dementia. Neurology. 2000;54:1304–1309.
11. O’Sullivan M, Ngo E, Viswanathan A, Jouvent E, Gschwendtner A,
Saemann PG, et al. Hippocampal volume is an independent predictor of
cognitive performance in CADASIL. Neurobiol Aging. 2009;30:
890 – 897.
12. Price JL, Ko AI, Wade MJ, Tsou SK, McKeel DW, Morris JC. Neuron
number in the entorhinal cortex and CA1 in preclinical Alzheimer
disease. Arch Neurol. 2001;58:1395–1402.
13. Kril JJ, Hodges J, Halliday G. Relationship between hippocampal volume
and CA1 neuron loss in brains of humans with and without Alzheimer’s
disease. Neurosci Lett. 2004;361:9–12.
14. Schmidt-Kastner R, Freund TF. Selective vulnerability of the hip-
pocampus in brain ischemia. Neuroscience. 1991;40:599– 636.
15. Wu W, Brickman AM, Luchsinger J, Ferrazzano P, Pichiule P, Yoshita
M, et al. The brain in the age of old: the hippocampal formation is
targeted differentially by diseases of late life. Ann Neurol. 2008;64:
16. Apostolova LG, Green AE, Babakchanian S, Hwang KS, Chou Y-Y,
Toga AW, et al. Hippocampal atrophy and ventricular enlargement in
normal aging, mild cognitive impairment (MCI), and Alzheimer disease.
Alzheimer Dis Assoc Disord. Epub ahead of print March 25, 2011. DOI:
17. West MJ, Coleman PD, Flood DG, Troncoso JC. Differences in the
pattern of hippocampal neuronal loss in normal ageing and Alzheimer’s
disease. Lancet. 1994;344:769–772.
18. Pakkenberg B, Pelvig D, Marner L, Bundgaard MJ, Gundersen HJ,
Nyengaard JR, et al. Aging and the human neocortex. Exp Gerontol.
19. Rossler M, Zarski R, Bohl J, Ohm TG. Stage-dependent and sector-
specific neuronal loss in hippocampus during Alzheimer’s disease. Acta
Gemmell et al Neuronal Atrophy and Poststroke Dementia 813
20. Korbo L, Amrein I, Lipp HP, Wolfer D, Regeur L, Oster S, et al. No
evidence for loss of hippocampal neurons in non-Alzheimer dementia
patients. Acta Neurol Scand. 2004;109:132–139.
21. Zarow C, Vinters HV, Ellis WG, Weiner MW, Mungas D, White L, et al.
Correlates of hippocampal neuron number in Alzheimer’s disease and
ischemic vascular dementia. Ann Neurol. 2005;57:896–903.
22. Terry RD, DeTeresa R, Hansen LA. Neocortical cell counts in normal
human adult aging. Ann Neurol. 1987;21:530–539.
23. Swaab DF, Hofman MA, Lucassen PJ, Salehi A, Uylings HB. Neuronal
atrophy, not cell death, is the main hallmark of Alzheimer’s disease.
Neurobiol Aging. 1994;15:369 –371; discussion 379–380.
24. Morrison JH, Hof PR. Selective vulnerability of corticocortical and hip-
pocampal circuits in aging and Alzheimer’s disease. Prog Brain Res.
25. Fjell AM, Walhovd KB. Structural brain changes in aging: courses,
causes and cognitive consequences. Rev Neurosci. 2010;21:187–221.
26. Freeman SH, Kandel R, Cruz L, Rozkalne A, Newell K, Frosch MP, et al.
Preservation of neuronal number despite age-related cortical brain
atrophy in elderly subjects without Alzheimer disease. J Neuropathol Exp
27. Ballard C, Rowan E, Stephens S, Kalaria R, Kenny RA. Prospective
follow-up study between 3 and 15 months after stroke: improvements and
decline in cognitive function among dementia-free stroke survivors ⬎75
years of age. Stroke. 2003;34:2440–2444.
28. Kalaria RN, Kenny RA, Ballard CG, Perry R, Ince P, Polvikoski T.
Towards defining the neuropathological substrates of vascular dementia.
J Neurol Sci. 2004;226:75–80.
29. Allan LM, Rowan EN, Firbank MJ, Thomas AT, Parry S, Polvikoski TM,
et al. Long term incidence of dementia, predictors of mortality and
pathological diagnosis in older stroke survivors. Brain. In press.
30. Perry RH, Oakley AE. Newcastle brain map. In Roberts G, Leigh P,
Weinberger D, eds. Neuropsychiatric Disorders.1
Ed. London, Wolfe
31. Insausti R, Juottonen K, Soininen H, Insausti AM, Partanen K, Vainio P,
et al. MR volumetric analysis of the human entorhinal, perirhinal, and
temporopolar cortices. AJNR Am J Neuroradiol. 1998;19:659– 671.
32. Canto CB, Wouterlood FG, Witter MP. What does the anatomical orga-
nization of the entorhinal cortex tell us? Neural Plast. 2008:381243,
33. Sterio DC. The unbiased estimation of number and sizes of arbitrary
particles using the dissector. J Microsc. 1984;134:127–136.
34. Rajkowska G, Miguel-Hidalgo JJ, Dubey P, Stockmeier CA, Krishnan
KR. Prominent reduction in pyramidal neurons density in the orbito-
frontal cortex of elderly depressed patients. Biol Psychiatry. 2005;58:
35. Gundersen HJ, Jensen EB. The efficiency of systematic sampling in
stereology and its prediction. J Microsc. 1987;147:229–263.
36. Gundersen HJ. The nucleator. J Microsc. 1988;151:3–21.
37. Schonheit B, Zarski R, Ohm TG. Spatial and temporal relationships
between plaques and tangles in Alzheimer-pathology. Neurobiol Aging.
38. Chevaleyre V, Siegelbaum SA. Strong CA2 pyramidal neuron synapses
define a powerful disynaptic cortico-hippocampal loop. Neuron. 2010;
39. Apostolova LG, Mosconi L, Thompson PM, Green AE, Hwang KS,
Ramirez A, et al. Subregional hippocampal atrophy predicts Alzheimer’s
dementia in the cognitively normal. Neurobiol Aging. 2010;31:
40. Harrison PJ, Eastwood SL. Neuropathological studies of synaptic con-
nectivity in the hippocampal formation in schizophrenia. Hippocampus.
41. van de Nes JA, Nafe R, Schlote W. Non-tau based neuronal degeneration
in Alzheimer’s disease—an immunocytochemical and quantitative study
in the supragranular layers of the middle temporal neocortex. Brain Res.
42. Gundersen HJ, Bagger P, Bendtsen TF, Evans SM, Korbo L, Marcussen
N, et al. The new stereological tools: dissector, fractionator, nucleator and
point sampled intercepts and their use in pathological research and
diagnosis. APMIS. 1988;96:857–881.
43. Firbank MJ, He J, Blamire AM, Singh B, Danson P, Kalaria RN, et al.
Cerebral blood flow by arterial spin labeling in poststroke dementia.
44. Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA,
Markesbery WR. Brain infarction and the clinical expression of Alzhei-
mer disease. The Nun Study. JAMA. 1997;277:813–817.
814 Stroke March 2012