Content uploaded by Lisa Cipolotti
Author content
All content in this area was uploaded by Lisa Cipolotti on Dec 13, 2013
Content may be subject to copyright.
Neurocase (2005) 11, 56–64
Copyright © Taylor & Francis Inc.
ISSN: 1355-4795 print
DOI: 10.1080/13554790490896866
Neurocase
A decade of pre-diagnostic assessment in a case of familial
Alzheimer’s disease: tracking progression from asymptomatic
to MCI and dementia
A.K. GODBOLT
1
, L. CIPOLOTTI
2
, V.M. ANDERSON
1
, H. ARCHER
1
, J.C. JANSSEN
1
, S. PRICE
1
, M.N. ROSSOR
1
,
and N.C. FOX
1
1
Dementia Research Group, Institute of Neurology, Queen Square, London and
2
Department of Neuropsychology, National Hospital
of Neurology and Neurosurgery, Queen Square, London
Detailed study of the very earliest phases of Alzheimer’s disease (AD) is seldom possible, especially those changes preceding the
development of mild cognitive impairment (MCI), which may occur years before diagnosis. Knowledge of imaging and neuropsychological
features of these early stages would add insight into this poorly understood phase of the disease. We present data from a subject who
entered a longitudinal study of individuals at risk of familial Alzheimer’s disease (FAD), as a healthy volunteer with no memory
complaints, undergoing 12 assessments between 1992 and 2003. Longitudinal MRI, neuropsychological and clinical data are presented
over the decade preceding this man’s diagnosis, through the asymptomatic and prodromal preludes to his presentation with MCI and on to
eventual conversion to AD.
Introduction
The concept of mild cognitive impairment (MCI) has increas-
ingly been adopted in recent years as the importance of iden-
tifying and understanding the earliest clinical phase of
Alzheimer’s disease (AD) has been recognized. Different
diagnostic criteria for MCI vary in their detail (Jack et al.,
1999; Petersen et al., 2001), but generally require memory
complaints from the patient and an informant, together with
objective evidence of memory impairment and generally nor-
mal cognition. Most MCI studies recruit patients who have
sought medical attention because of memory complaints, pre-
sumably because these are of a degree to either concern the
patient or impact their daily functioning, even if this is to a
much lesser extent than in those receiving a diagnosis of
dementia. Whilst the rate of conversion to AD is undoubtedly
considerably raised in patients with MCI, it is not 100%, and
those with MCI comprise a rather heterogeneous group,
including nevertheless a significant proportion that do subse-
quently develop AD.
The pathological changes underlying AD are likely to start
some time before even mild symptoms may be present, and it
has not generally been possible to study this phase of the dis-
ease in any detail. The occurrence of AD in a familial form
(FAD) with autosomal dominant inheritance provides an
opportunity to study this very earliest phase of the disease.
Individuals at 50% risk of developing FAD can be recruited
when entirely well and studied serially in detail in the years
before the expected age at onset in their family, with the pre-
diction that half of the cohort would go on to develop AD.
This paradigm has been successfully used in several previous
studies of subjects at risk of FAD (Kennedy et al., 1995; Fox
et al., 1998; Fox et al., 2001).
We present data on one subject who presented to medical
services and received a diagnosis of FAD nearly 11 years
after first entering a longitudinal imaging and neuropsycho-
logical study. This very long period of prediagnostic assess-
ment allows insight into the early phases of AD.
Case report
This man, TOP (not the subject’s initials), entered a longitu-
dinal study of individuals at 50% risk of familial Alzheimer’s
disease (FAD) in 1992, at the age of 50 years. This study was
approved by the local ethics committee and informed consent
was obtained from TOP.
Received 28 January, 2004; accepted 31 August 2004
We thank TOP and his wife for participating in this study, and the
assistants of the neuropsychology department of the National Hos-
pital of Neurology and Neurosurgery who performed the neuro-
psychology assessments and assisted in data collation, including
Ms Susanna Cole. Mutation analysis was performed by Professor
John Collinge and Mr John Beck of the MRC Prion Unit, Institute
of Neurology. Funding was provided by an MRC programme grant,
number G9626876. NCF holds an MRC senior clinical scientist
fellowship.
Address correspondence to N.C. Fox, Institute of Neurology
Dementia Research Group, Queen Square, London WCIN 3BG,
United Kingdom. E-mail: n.fox@dementia.ion.ucl.ac.uk
Pre-diagnostic assessment in a case of FAD 57
At the time of study entry TOP had no cognitive com-
plaints, and his asymptomatic status was confirmed indepen-
dently by his wife. He is a member of a British FAD pedigree
in which affected members carry the amyloid precursor pro-
tein (APP) V717I mutation. The mean age at onset of FAD in
his family is 53 years (range 43 to 58 years). His genetic
status was unknown until 2003 when diagnostic genotyping
was performed as part of investigation of his memory com-
plaints. He was found to carry the APP V717I mutation and
to have ApoE ε3, ε3 alleles.
According to the longitudinal study protocol, he underwent
serial assessments at approximately yearly intervals. Each
assessment comprised brain magnetic resonance imaging
(MRI), neuropsychology assessment and medical assessment
including mini mental state examination (MMSE) and clini-
cal dementia rating (CDR) according to a standard protocol,
as previously reported (Fox et al., 1996; Fox et al., 1998).
Assessments were performed by a clinical research fellow in
neurology and a neuropsychologist or assistant under super-
vision, who were blinded to the outcome of assessments other
than their own.
As part of the study protocol, TOP and his wife were made
aware that results would not be available from the research
assessments, but that should they become concerned about
symptoms, a clinical assessment would be arranged at their
request. TOP did not seek such an assessment until late
2002.
Methods
Image acquisition and analysis
MRI brain scans were acquired on two 1.5 Tesla scanners:
one from 1992 until 1999 and the other from 2000 until 2003.
A routine sagittal (T1 weighted) scout sequence, an axial
dual-echo sequence (T2 and proton density weighted) and a
T1 weighted volumetric image (performed in the coronal
plane using a spoiled gradient echo technique with a 24cm
field of view yielding 1.5mm thick contiguous slices on a
256x128 image matrix) were obtained at each assessment.
Follow-up scans were accurately registered onto baseline
images, using a nine degrees of freedom rigid body registra-
tion and rates of whole brain atrophy were calculated using
the Brain Boundary Shift Integral (BBSI), (Freeborough and
Fox, 1997). Hippocampal and ventricular volumes were mea-
sured manually as previously described (Scahill et al., 2003).
Neuropsychology assessments
Neuropsychological tests assessed intellectual functioning,
current (Weschsler Adult Intelligence Scale Revised)
(Weschler, 1981), and premorbid (National Adult Reading
Test (NART)) (Nelson and Willison, 1991); verbal and visual
recognition memory (Recognition Memory Test (RMW and
RMF)) (Warrington, 1984); verbal recall memory (Paired
Associate Learning (PAL)) (Warrington, 1996); topographi-
cal memory (Topographical Memory Recognition Test
(CTRMT)) (Warrington, 1996); naming (Graded Naming
Test (GNT)) (McKenna and Warrington, 1983); spelling
(Oral Graded Difficulty Spelling Test (GDST)) (Baxter and
Warrington, 1994); calculation (Graded Difficulty Arithmetic
Test (GDAT)) (Jackson and Warrington, 1986); and visuo-
perceptual abilities (Silhouettes (Sil) from the Visual Object
and Space Perception Battery) (Warrington and James,
1991). Executive function was assessed with the modified
Wisconsin Card Sorting Test (MCST) (Nelson, 1976) and
Cognitive Estimates (Shallice and Evans, 1978). Parallel ver-
sions were available for the RMW and RMF and were used in
rotation.
Three derived scores were calculated as follows:
1. The difference between the verbal and performance scale
IQ and the NART was calculated to compare current and
baseline intellectual function.
2. The memory, naming, spelling, calculation and visual per-
ceptual scores were derived by converting the standard-
ized test performance into percentile scores, according to
normative data.
3. For the frontal tasks, percentile scores could not be
derived as their scores are not normally distributed. Thus,
the following two procedures were used: a) the responses
on the modified Wisconsin Card Sorting Test were ana-
lyzed in terms of number of categories and number of total
errors; b) a pass or fail procedure was adopted for the Cog-
nitive Estimates test.
The diagnosis of Mild Cognitive Impairment was made
retrospectively, according to the criteria published in 1999
(Jack et al., 1999):
1. Memory complaint documented by the patient or collateral
source
2. Normal general cognitive function
3. Normal activities of daily living
4. No dementia
5. Objective memory impairment, defined by performance at
1.5 standard deviations below age and education-matched
controls on indices of memory (RMT used for this criteria).
6. Clinical Dementia Rating scale score of 0.5
Their other requirement was for age range of 60–89 years.
This criterion was not included due to the early onset familial
nature of the disease in TOP.
The diagnosis of probable AD was made according to the
National Institute of Neurologic and Communicative Dis-
orders and Stroke/AD and Related Disorders Association
(NINCDS/ADRDA) criteria for AD (McKhann et al.,
1984).
58 Godbolt et al.
History
TOP is right-handed and was a self-employed rural craftsman,
having completed an apprenticeship after leaving school at
the age of 15 years without qualifications. A summary of his
assessments is given in Table 1. At the time of study entry in
1992 he had no past medical history of note and was an ex-
smoker who drank four units of alcohol per day. He scored
30/30 on the MMSE. Neurological examination was normal.
At each visit a history was taken from the patient and his
wife independently. From 1992 until 1996 both the patient
and his wife considered him to be asymptomatic. The first
suggestion of cognitive problems was noted in 1997, at the
age of 55 years. In response to questions about his memory,
his wife commented that he had trouble passing on messages
and sometimes forgot she had told him things. She thought
he was no worse than an unrelated friend of the same age,
but did think his memory was a little worse than before.
TOP himself said he had no problems at work but when
asked about his memory, did think that this was slightly
worse than previously. His work involved complex route
finding on poorly marked roads, and although he had not
become lost he volunteered that he was having to think more
about which turns to take. He was also misplacing objects
and had missed occasional appointments. However, he contin-
ued to perform well in the local card playing league, ending
in the top quarter.
However, it was not until late 2002 that TOP sought medical
advice regarding his memory difficulties and was referred for
assessment, receiving a clinical diagnosis of familial Alzhe-
imer’s disease in early 2003. At the time of referral he was
still working. Of interest, at the time of his diagnostic assess-
ment in early 2003 TOP and his wife dated his cognitive diffi-
culties back only two to three years, despite actually having
commented on a worsening of his memory six years earlier.
He was assessed on five occasions between first express-
ing concerns about his memory in 1997 and a diagnosis of
Alzheimer’s disease being made in 2003. In 1998 his wife
was less concerned about his memory, and he reported no
progression in his difficulties. In 1999 his wife reported that
he had no problems at work but minor difficulties at home
continued. However, she did not think these were serious and
attributed them to stressful family events. TOP felt he was
more reliant on his diary and on planning routes, but was also
anxious about family events. In 2000 he again reported diffi-
culties determining the best route to visit clients, but
described himself as “quick witted” and felt his memory was
unchanged. His wife reported that memory difficulties con-
tinued, misplacing items at home, though he had also
resumed an old hobby successfully. He was found to be
hypertensive in the months prior to this visit and had com-
menced beta blockers. In 2001 he was taking the wrong turn-
ing more frequently, and had difficulty remembering
directions to new clients. His wife reported increasing diffi-
culties with memory and route finding. By 2002 his wife was
more certain that memory problems had progressed. He was
repetitive in conversation and sometimes lost track of what
he was saying. He was still working, though had less work
than previously, and sometimes couldn’t find the routes to his
clients. He had had occasional “near misses” when driving
and had difficulties cooking meals when his wife was unable
to do this because of illness. TOP was aware that his memory
was worsening and again reported difficulties with route find-
ing. He sought medical attention a few months after this
assessment.
Following the diagnosis of Alzheimer’s disease TOP was
advised not to drive and retired from work. Prior to his last
assessment Rivastigmine was commenced with some reported
benefit. His wife reported that he was unable to use public
transport alone, had difficulty finding a nearby relative’s
T
able 1. Summary of serial assessments for TOP
E
vent Year
Assessment
number
Age at
assessment (years)
Years from
symptom onset
Years from
diagnosis of AD MMSE CD
R
S
tudy entry 1992 1 50.5 −4.7 −10.8 30 0
1994 2 52.3 −3.1 −9.0 26 0
1995 3 53.5 −1.9 −7.8 29 0
1996 4 54.5 −0.9 −6.8 28 0
S
ymptom onset 1997 5 55.3 0.0 −5.9 27 0
1998 6 56.3 0.9 −5.0 29 0
1999 7 57.2 1.9 −4.0 27 0
2000 8 58.3 2.9 −3.0 28 0
2001 9 59.6 4.2 −1.7 26 0
F
irst sought medical
attention
2002 10 60.7 5.3 −0.8 24 0.5
F
irst met criteria
for AD
Mar 2003 11 61.2 5.9 0.0 20 1
R
ivastigmine commenced
prior to this assessment
Sep 2003 12 61.8 6.5 0.6 22 1
Pre-diagnostic assessment in a case of FAD 59
house and had more trouble following a longstanding hobby.
He needed prompting to wear clean clothes. However he
managed well when left alone, was able to buy items from
the local shop, and did not become lost in his immediate
locality.
MMSE scores are shown in Figure 1. These were highly
correlated with time from diagnosis of AD (r
2
=0.61,
p=0.0026).
Neuropsychological profile
Neuropsychological findings are summarized in Table 2.
Almost all cognitive domains were first evaluated 10.8 years
before diagnosis. The only exceptions were the PAL test of
verbal recall, first administered 9 years before diagnosis, and
a stringent visual recognition memory test (CTRMT), first
administered 5 years before diagnosis.
At first assessment TOP had average intellectual function.
Visual and verbal recognition memory scored above the
75th%, whilst verbal recall was initially weaker, scoring at
the 10–25th%. Naming was normal (GNT 50–75th%), as were
spelling (GDST 10–25th%), calculation (GDAT 50–75th%),
visuoperceptual function (Silhouettes 25–50th%) and frontal
executive function (MCST).
Practice effects were present throughout the first few
sequential assessments. For example, full IQ initially increased
over several years, with a peak performance IQ 27 points
above baseline (achieved at his sixth assessment, after
5.8 years), and peak verbal IQ of 15 points above baseline (at
his eighth assessment, after 7.8 years). Verbal recall, weak at
the first assessment, rose to above the 75th% with practice
at the third assessment.
At the final assessment, seven months after the clinical
diagnosis of AD was made (and approximately 11 years after
the first assessment), all cognitive domains were reassessed.
His scores of intellectual function remained in the average
range. However, this represented a considerable decline of
33 points from his peak performance IQ achieved five years
before diagnosis of AD, and a decline of 17 points from his
peak verbal IQ, achieved three years before diagnosis (see
Figure 2). TOP’s verbal memory scores (recall and recogni-
tion) were at the 5
th
–10
th
% and his topographical recognition
memory score below the 5th%. His performance on the visual
recognition memory test was weak, at the 10–25th%. He had
commenced Rivastigmine three months prior to this final assess-
ment. His recognition memory scores showed a trend to decline
from many years before diagnosis (see Figure 3), (RMW
r
2
=0.81, p=0.0001; RMF r
2
=0.44, p=0.018), though it was only
at the assessment nine months before the diagnosis of AD that
they were clearly poor in relation to normative data (greater than
1.5SD below the mean, and below the 5th%). Indeed stable or,
more likely, improving scores would be expected with sequen-
tial assessment if no decline was taking place.
At final assessment TOP scored at the 5–10th% on the first
trial of the PAL test of verbal recall, and at the 10-25th% on
the second trial (see Figure 4a). Although previous scores are
higher, gradual decline is seen as he becomes symptomatic,
almost six years before diagnosis on Trial 1 and two years
before diagnosis on Trial 2. Static topographical memory
(CTRMT) was first assessed five years before diagnosis,
when he scored above the 75th%, and became impaired at
least nine months before diagnosis, when TOP first met criteria
for MCI (see Figure 4b).
Other features by final assessment were as follows. Naming
scored above the 75th% even at the last assessment. Calcula-
tion declined from 75th% to 10–25th%, and spelling from
25–50th% to 5–10th%, both after an initial improvement due
to practice. Perception scores remained good at 50–75th% at
the final assessment. However, again, the practice effect seen
until three years before diagnosis was then lost. Scores for
naming, spelling, calculation and perception are shown in
Figure 5. Executive dysfunction was present at final assess-
ment, achieving only two categories on the MCST and making
many errors (Figure 6), although TOP passed the cognitive
estimates test at every time point
Imaging data
There was no evidence of significant ischaemic change on
any scans, as illustrated by TOP’s final scan (Figure 7). Sub-
tle global and hippocampal atrophy is apparent from regis-
tered coronal images acquired over the decade preceding
diagnosis of AD in TOP (Figure 8), and is more obvious by
the time of diagnosis. However, in isolation all but the most
recent two scans would be considered normal on visual
assessment of the images.
Global atrophy rates, calculated using the brain boundary
shift integral, are shown in Figure 9. Rates shown are in rela-
tion to baseline scans on two different scanners: one until
four years before diagnosis of AD and another from three
years before diagnosis of AD. This avoids the inaccuracies
inherent in cross scanner registration. Mean rates of atrophy
of 0.32% per year (95% CI 0.10 – 0.54%) are found in con-
trol subjects (Scahill et al., 2003), and of 2.37% per year (SD
1.11%) in subjects with AD (Fox et al., 2000).
F
ig. 1. MMSE scores for TOP. Data are fitted with a linear tren
d
l
ine. The time at which TOP fulfilled objective criteria for MCI
is
i
ndicated by a dashed line.
60
Table 2. Summary of neuropsychological profile for TOP. Percentile scores are given in each domain at initial assessment, peak percentile band, and final assessment, together
with time of the beginning of a downward trend in his performance for each domain. See text for abbreviations
Initial
assessment
time
Initial
assessment
percentile band
Time highest
percentile band
first achieved
Peak
percentile
band
Beginning of
downward trend
in performance
Final
assessment
time
Final
assessment
percentile band
Test Years from
diagnosis
of AD
Years from
diagnosis
of AD
Years from
diagnosis
of AD
Years from
diagnosis
of AD
Intellectual
function
VIQ −10.8 502 <75 −4.0 >75 −3.0 0.6 252 50
PIQ −10.8 502 <75 −9.0 >75 −5.0 0.6 502 75
Memory RMW −10.8 >75 −10.8 >75 −10.8 0.6 52 <10
RMF −10.8 >75 −10.8 >75 −10.8 0.6 102 <25
PAL −9.0 102 <25 −7.8 >75 −6.8 (trial 1) 0.6 52 <10
CTRMT −5.0 >75 −5.0 >75 −4.0 0.6 12 <5
Naming GNT −10.8 502 <75 −9.0 >75 −1.7 0.6 >75
Calculation GDAT −10.8 502 <75 −9.0 >75 −5.8 0.6 102 <25
Spelling GDST −10.8 102 <25 −9.0 252 <50 −4.0 0.6 52 <10
Visuoperceptual
skills
Sil −10.8 252 <50 −5.0 >75 −1.7 0.6 502 <75
Number Number Number
Executive
function
MCST
categories
−10.8 6 −10.8 6 0.6 0.6 2
MCST
errors
−10.8 3 −7.8 2 −0.8 0.6 23
Pre-diagnostic assessment in a case of FAD 61
F
ig. 2. Performance and verbal IQ related to baseline NART. T
he
t
ime at which TOP fulfilled objective criteria for MCI is indicat
ed
b
y a dashed line.
F
ig. 3. Recognition memory scores fitted with linear trend line
s.
T
he time at which TOP fulfilled objective criteria for MCI
is
i
ndicated by a dashed line.
F
ig. 4. (a) Verbal recall scores. The time at which TOP fulfill
ed
o
bjective criteria for MCI is indicated by a dashed line.
(
b) Topographical memory scores. The time at which TOP fulfill
ed
o
bjective criteria for MCI is indicated by the dashed line.
F
ig. 5. Raw scores for naming, spelling, calculation and perceptio
n.
T
he time at which TOP fulfilled objective criteria for MCI
is
i
ndicated by a dashed line.
F
ig. 6. Executive function. The time at which TOP fulfill
ed
o
bjective criteria for MCI is indicated by a dashed line.
F
ig. 7. Axial T2 weighted image from final assessment showing n
o
e
vidence of significant ischaemic change.
62 Godbolt et al.
Hippocampal volumes are shown in Figure 10. A decline
in hippocampal volume is demonstrated, from nine years
before diagnosis on the right and from six years before diag-
nosis on the left. Ventricular volumes are shown in Figure 11.
The rate of ventricular expansion appears to accelerate from
five years before diagnosis. Data are shown with fitted third
order polynomial curves, based on the following: Regression
analysis supported non-linearity (right hippocampus p=0.001,
left hippocampus p=0.08, ventricles p<0.001), and gave some
support for a 3rd order polynomial being a better fit than a
2nd order polynomial for the right hippocampus and ventri-
cles (right hippocampus p=0.06, left hippocampus p=0.53,
ventricles p=0.007). A 3rd order polynomial fit was also
more biologically plausible; a 2nd order polynomial suggest-
ing a biologically implausible initial increase in hippocampal
volume and decrease in ventricular volume.
Discussion
This study gives further insight into the earliest stages of AD.
The extensive clinical, imaging and neuropsychological data
collected on patient TOP allow us to document the long pro-
drome of AD. Moreover, they allow monitoring of the very
gradual nature of early cognitive and neuroanatomical
changes as well as highlighting methodological issues inher-
ent in longitudinal data acquisition and analysis.
In this case we were able to make the diagnosis of AD with
confidence even though TOP remains mildly affected, given
the clinical features and detailed neuropsychology and imag-
ing data, together with the APP mutation finding. There was
no significant co-morbidity. Despite his mild hypertension no
significant cerebrovascular disease was demonstrated by T2
imaging. There was no evidence of depressive illness predi-
agnostically. The blinding of the assessors to his mutation
status prior to diagnosis strengthens the validity of the previ-
ous assessments.
Overall, the neuropsychological assessments show evi-
dence of decline in several domains over the decade before
diagnosis. However, it was evident his scores on selective
cognitive tests showed some variability. The recognition that
fluctuations in scores and practice effects occur has recently
led to attempts to separate these from real change, with the
recent investigation of reliable change indices for the RMT
(Bird et al., 2003), the Graded Naming Test (GNT), the Sil-
houettes Test and the Modified Wisconsin Card Sorting Test
(Bird et al., 2004). However, these evaluations involved only
one repeat of the test over a one-month interval. Therefore
they were considered an inappropriate method of analysis for
this study. The most appropriate control data would be results
from at-risk subjects similarly assessed over several years.
F
ig. 8. Sequential registered coronal T1 MRI brain slices for TOP
.
F
ig. 9. Whole brain atrophy. The time at which TOP fulfill
ed
o
bjective criteria for MCI is indicated by a dashed line.
F
ig. 10. Left and right hippocampal volumes. Data are shown wi
th
f
itted third order polynomial curves. The time at which TO
P
f
ulfilled objective criteria for MCI is indicated by a dashed line.
F
ig. 11. Ventricular volumes. Data are shown with a fitted thi
rd
o
rder polynomial curve. The time at which TOP fulfilled objecti
ve
c
riteria for MCI is indicated by a dashed line.
Pre-diagnostic assessment in a case of FAD 63
The collection of these data are ongoing and when available
will allow appropriate analysis of practice effects over this
long period of assessment.
In several domains (intellectual function, calculation,
spelling, visuoperceptual function) scores initially rose with
practice and then fell gradually as the earliest sign of cogni-
tive dysfunction. The timing of this fall varied with domain,
beginning five years before diagnosis of AD for performance
IQ, three years before diagnosis for verbal IQ, and six and
four years before diagnosis for calculation and spelling,
respectively. It is only possible to draw such conclusions
given the longitudinal nature of the data.
TOP’s first symptoms suggested possible early impairment
of topographical memory. He indeed complained of a feeling
of difficulty with complex route finding on unmarked country
roads. Topographical recognition memory was not formally
assessed before he first reported these symptoms. Interest-
ingly, topographical recognition memory scores were initially
normal and started to decline only several years after his first
reported symptoms. More stringent tests tapping both allo-
centric as well as egocentric spatial memory may be required
to detect the very earliest stages of topographical dysfunction
in AD.
The imaging data demonstrate the onset of pathology many
years before diagnosis of AD. Global atrophy exceeded that
of controls around three years before diagnosis. Decline in
hippocampal volume began several years earlier, and acceler-
ation in ventricular expansion is seen from about five years
before diagnosis. Imaging changes begin at about the time of
earliest symptoms, though it is important to note that these
symptoms were subtle, elicited on direct questioning about
memory rather than spontaneously brought to medical atten-
tion, and preceded fulfilment of MCI criteria by several
years. The finding of changes in both neuropsychological and
imaging data years before diagnosis fit with pathological data
suggesting that AD changes begin years before diagnosis
(Riley et al., 2002).
Due to practice effects on the neuropsychological assess-
ments, it is possible that objective neuropsychological evi-
dence for memory impairment in relation to normative data,
as required by the criteria chosen for MCI, was demonstrated
later than if serial assessments had not been performed. This
may possibly also have led to underestimation of the rate of
progression of cognitive impairment. Nevertheless, the start
of symptoms some six years before the diagnosis of AD sug-
gests a very long prodromal phase, and emphasises that esti-
mates of disease duration, usually beginning at diagnosis
(Brookmeyer et al., 2002), may ignore a substantial part of
the disease course.
In previous studies of subjects at risk of familial AD who
subsequently developed clinically diagnosed AD, the most
common symptoms declared were very mild episodic memory
problems (Fox et al., 1996). Such subjects may be especially
alert to early symptoms due to anxiety about their family
history. Hence TOP’s first symptoms in 1997 probably repre-
sent a much earlier stage in the disease than the point at
which first symptoms are noted in sporadic AD. In 1997
TOP’s difficulties were very subtle, and of an insufficient
degree to meet criteria subsequently published for MCI (Jack
et al., 1999). The difficulty in dating onset is illustrated in
this case by the discrepancy between TOP and his wife’s con-
temporaneous dating of cognitive symptoms in 1997 (nearly
six years before diagnosis) when specifically asked about the
presence of any symptoms, and the history they gave at the
time of the diagnostic visit in 2003 (of two to three years of
cognitive decline).
Indeed, the earliest indication of Alzheimer’s disease may
be at a time when patients first develop subtle changes in
their memory, without objective evidence of memory impair-
ment on neuropsychological tests, and who will not initially
meet criteria for diagnosis of MCI (Jack et al., 1999). Data
from population studies exploring the relationship between
memory complaints and the development of dementia are
inconclusive. However, a meta-analysis of these studies
carried out by Jonker et al. (2000) emphasized that subjective
symptoms of memory loss should be taken seriously in high
functioning individuals and the elderly, even in the absence
of objective memory deficits. This earliest disease stage is
difficult to study, as practical considerations preclude the
detailed neuropsychological and imaging study of large num-
bers of healthy elderly subjects for the prolonged periods
needed, in order to yield data many years before diagnosis in
only a few subjects.
A decision about whether a patient with subtle cognitive
complaints has MCI or not can be difficult, and even appar-
ently detailed criteria can be difficult to apply in individual
cases. For example, MCI criteria generally state that there
should be objective evidence of memory impairment, and
one study of MCI (Jack et al., 1999) defined this as scores
at least 1.5 SD below the population mean on memory
tests. The patient who scores within the normal range on
some tests and just below this on others is difficult to
classify, as it could be argued that there is also objective
evidence of normal memory; whilst with extensive testing
even healthy controls will perform poorly on some tests.
TOP did not meet all of the criteria for MCI (Jack et al.,
1999) until 9 months before diagnosis of AD. However, if
one considers, for example, his recognition memory scores
in isolation, he did drop 1.5SD below the age-matched
mean for the RMW three years before diagnosis, although
at this time his CDR was 0, and his RMF score was above
the 1.5SD cut-off.
Decline is a continuous process, and whilst classification
of cases as normal, MCI and AD based on a single assess-
ment is necessary for clinical purposes, and is useful in mak-
ing group comparisons, it may be less useful in individual
cases than a consideration of the rate and nature of longitudi-
nal change. Without the benefit of serial assessments, deci-
sions about whether an individual is in the earliest stages of
Alzheimer’s disease inevitably rest on an estimate of that
individual’s previous cognitive function with a history of
decline provided by the subject and a close informant, or
64 Godbolt et al.
from population means (hence the criteria for MCI). This
case further illustrates the very gradual nature of the earliest
changes and how difficult it is to pinpoint the onset of patho-
logical cognitive change.
These same considerations apply to the detection of patho-
logical structural change on MRI in AD. Rates of decline
appear to increase many years before diagnosis at a point
when individual measures of brain, hippocampal and ventricular
volume are within the normal range. Both memory impairment
and medial temporal lobe atrophy seem to occur early. More
sensitive measures of both of these modalities are needed to
detect the very earliest changes. Serial assessment has a valuable
role in assessing these changes.
References
Baxter DM, Warrington EK. Measuring dysgraphia: A graded-difficulty
spelling test. Behavioural Neurology 1994; 7: 107–116.
Bird CM, Papadopoulou K, Ricciardelli P, Rossor MN, Cipolotti L. Test-
retest reliability, practice effects and reliable change indices for the
recognition memory test. Br.J Clin Psychol. 2003; 42: 407–425.
Bird CM, Papadopoulou K, Ricciardelli P, Rossor MN, Cipolotti L. Moni-
toring cognitive changes: psychometric properties of six cognitive
tests. Br.J Clin Psychol. 2004; 43: 197–210.
Brookmeyer R, Corrada MM, Curriero FC, Kawas C. Survival following a
diagnosis of Alzheimer disease. Arch.Neurol. 2002; 59: 1764–1767.
Fox NC, Cousens S, Scahill R, Harvey RJ, Rossor MN. Using serial registered
brain magnetic resonance imaging to measure disease progression in
Alzheimer disease: Power calculations and estimates of sample size to
detect treatment effects [see comments]. Arch Neurol 2000; 57: 339–344.
Fox NC, Crum WR, Scahill RI, Stevens JM, Janssen JC, Rossor MN.
Imaging of onset and progression of Alzheimer’s disease with voxel-
compression mapping of serial magnetic resonance images. Lancet
2001; 358: 201–205.
Fox NC, Warrington EK, Freeborough PA, Hartikainen P, Kennedy AM,
Stevens JM, Rossor MN. Presymptomatic hippocampal atrophy in
Alzheimer’s disease: A longitudinal MRI study. Brain 1996; 119:
2001–2007.
Fox NC, Warrington EK, Seiffer AS, Agnew SK, Rossor MN. Presy-
mptomatic cognitive deficits in individuals at risk of familial
Alzheimer’s disease. A longitudinal prospective study. Brain 1998;
121: 1631–1639.
Freeborough PA, Fox NC. The boundary shift integral: an accurate and
robust measure of cerebral volume changes from registered repeat
MRI. Ieee Transactions On Medical Imaging 1997; 16: 623–629.
Jack CR, Petersen RC, Xu YC, Obrien PC, Smith GE, Ivnik RJ, Boeve BF,
Waring SC, Tangalos EG, Kokmen E. Prediction of AD with MRI-
based hippocampal volume in mild cognitive impairment. Neurol
1999; 52: 1397–1403.
Jackson M, Warrington EK. Arithmetic skills in patients with unilateral
cerebral lesions. Cortex. 1986; 22: 611–620.
Jonker C, Geerlings MI, Schmand B. Are memory complaints predictive for
dementia? A review of clinical and population-based studies 3. Int J
Geriatr Psychiatry 2000; 15: 983–991.
Kennedy AM, Frackowiak RSJ, Newman SK, Bloomfield PM, Seaward J,
Roques P, Lewington G, Cunningham VJ, Rossor MN. Deficits in
cerebral glucose metabolism demonstrated by positron emission
tomography in individuals at risk of familial Alzheimer’s disease.
Neurosci Lett 1995; 186: 17–20.
McKenna P, Warrington EK. The Graded Naming Test. Windsor, UK:
NFER-Nelson, 1983.
McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM.
Clinical diagnosis of Alzheimer’s Disease: Report of the NINCDS-
ADRDA work group under the auspices of Department of Health and
Human Services Task Force on Alzheimer’s Disease. Neurol 1984; 34:
939–944.
Nelson HE. A modified card sorting test sensitive to frontal lobe defects.
Cortex 1976; 12: 313–324.
Nelson HE, Willison J. The National Adult Reading Test (NART) manual.
Windsor, UK: NFER-Nelson, 1991.
Petersen RC, Stevens JC, Ganguli M, Tangalos EG, Cummings JL, DeKosky
ST. Practice parameter: Early detection of dementia: Mild cognitive
impairment (an evidence-based review). Neurol 2001; 56: 1133–1142.
Riley KP, Snowdon DA, Markesbery WR. Alzheimer’s neurofibrillary
pathology and the spectrum of cognitive function: findings from the
Nun Study. Ann Neurol 2002; 51: 567–577.
Scahill RI, Frost C, Jenkins R, Whitwell JL, Rossor MN, Fox NC. A lon-
gitudinal study of brain volume changes in normal aging using
serial registered magnetic resonance imaging. Arch.Neurol 2003;
60: 989–994.
Shallice T, Evans ME. The involvement of the frontal lobes in cognitive
estimation. Cortex 1978; 14: 294–303.
Warrington EK. Manual for the Recognition Memory Test for words and
faces. Windsor: NFER-Nelson, 1984.
Warrington EK. The Camden Memory Tests Manual. Hove: Psychology
Press, 1996.
Warrington EK, James M. The visual object and space perception battery.
Bury St Edmunds: Thames Valley Test Co, 1991.
Weschler D. Manual for the Wechsler Adult Intelligence Scale - Revised.
New York Psychological Corp, 1981.
