Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS-ADRDA criteria

/Home/Alzheimer's Diagnosis/Diagnostic Tests for Alzheimers/Alzheimer's Diagnosis Overviews

The NINCDS-ADRDA and the DSM-IV-TR criteria for Alzheimer's disease (AD) are the prevailing diagnostic standards in research; however, they have now fallen behind the unprecedented growth of scientific knowledge. Distinctive and reliable biomarkers of AD are now available through structural MRI, molecular neuroimaging with PET, and cerebrospinal fluid analyses. This progress provides the impetus for our proposal of revised diagnostic criteria for AD. Our framework was developed to capture both the earliest stages, before full-blown dementia, as well as the full spectrum of the illness. These new criteria are centred on a clinical core of early and significant episodic memory impairment. They stipulate that there must also be at least one or more abnormal biomarkers among structural neuroimaging with MRI, molecular neuroimaging with PET, and cerebrospinal fluid analysis of amyloid beta or tau proteins. The timeliness of these criteria is highlighted by the many drugs in development that are directed at changing pathogenesis, particularly at the production and clearance of amyloid beta as well as at the hyperphosphorylation state of tau. Validation studies in existing and prospective cohorts are needed to advance these criteria and optimise their sensitivity, specificity, and accuracy.


Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS-ADRDA criteria.
Dubois B, Feldman HH, Jacova C, Dekosky ST, Barberger-Gateau P, Cummings J, Delacourte A, Galasko D, Gauthier S, Jicha G, Meguro K, O'brien J, Pasquier F, Robert P, Rossor M, Salloway S, Stern Y, Visser PJ, Scheltens P.

INSERM U610, Hôpital de la Salpêtrière, Paris, France. bruno.dubois@psl.aphp.fr

BACKGROUND

For research purposes, the diagnosis of Alzheimer’s
disease (AD) is based on the criteria of the Diagnostic
and Statistical Manual of Mental Disorders, fourth
edition (DSM-IV-TR)[1] and the National Institute of
Neurological Disorders and Stroke–Alzheimer Disease
and Related Disorders (NINCDS–ADRDA) working
group.[2] These accepted criteria are fulfilled in a two-
step diagnostic process where there is initial
identification of a dementia syndrome and then the
application of criteria based on the clinical features of
the AD phenotype. The DSM-IV-TR criteria require the
presence of both a memory disorder and impairment in
at least one additional cognitive domain, both of which
interfere with social function or activities of daily living
(ADL).[1] ADL impairment has come to define the
threshold for the diagnosis of dementia beyond
the identification of a cognitive abnormality. The
NINCDS–ADRDA clinical criteria of probable AD do
not require evidence of interference with social or
occupational functioning but they include the
specification that the onset of AD is insidious and that
there is a lack of other systemic or brain diseases that
may account for the progressive memory and other
cognitive deficits. The currently accepted criteria
support a probabilistic diagnosis of AD within a clinical
context where there is no definitive diagnostic
biomarker. A definite diagnosis of AD is only made
according to the NINCDS–ADRDA criteria when
there is histopathological confirmation of the clinical
diagnosis.[2]

Since the publication of the NINCDS–ADRDA criteria
in 1984, the elucidation of the biological basis of AD has
advanced greatly, allowing an unprecedented under-
standing of the disease process. The clinical phenotype
of AD is no longer described in exclusionary terms, but
can be characterised more definitively on a phenotypic
basis. Distinctive markers of the disease are now
recognised including structural brain changes visible on
MRI with early and extensive involvement of the medial
temporal lobe (MTL), molecular neuroimaging changes
seen with PET with hypometabolism or hypoperfusion
in temporoparietal areas, and changes in cerebrospinal
fluid biomarkers. A driving force behind this emerging
identity of AD has been the intense research interest in
characterising the earliest stages of AD that predate the
crossing of the dementia threshold, defined by functional
disability. Prodromal AD (see glossary, panel 1) must be
distinguished within the broad and heterogeneous state
of cognitive functioning that falls outside normal ageing.[3]
This state has been described by a wide range of
nosological terms including age-associated memory
impairment, age-related cognitive decline, age-associated
cognitive decline, mild cognitive disorder, mild
neurocognitive disorder, cognitively impaired not
demented, and mild cognitive impairment.[4–9] Mild
cognitive impairment (panel 1) is the most widely used
diagnostic term for the disorder in individuals who have
subjective memory or cognitive symptoms, objective
memory or cognitive impairment, and whose activities
of daily living are generally normal. Progression to
clinically diagnosable dementia occurs at a higher rate
from mild cognitive impairment than from an
unimpaired state, but is clearly not the invariable clinical
outcome at follow-up.[10] A more refined definition of AD
is still needed to reliably identify the disease at its earliest
stages.


________
Panel 1: Glossary of terms

- Mild cognitive impairment

Variably defined but includes subjective memory or cognitive
symptoms or both, objective memory or cognitive
impairment or both, and generally unaffected activities of
daily living; affected people do not meet currently accepted
dementia or AD diagnostic criteria

- Amnestic mild cognitive impairment

A more specified term describing a subtype of mild cognitive
impairment, in which there are subjective memory symptoms
and objective memory impairment; other cognitive domains
and activities of daily living are generally unaffected; affected
people do not meet currently accepted dementia or AD
diagnostic criteria

- Preclinical AD

The long asymptomatic period between the first brain lesions
and the first appearance of symptoms and which concerns
normal individuals that later fulfil AD diagnostic criteria

- Prodromal AD

The symptomatic predementia phase of AD, generally
included in the mild cognitive impairment category; this
phase is characterised by symptoms not severe enough to
meet currently accepted diagnostic criteria for AD

- AD dementia

The phase of AD where symptoms are sufficiently severe to
meet currently accepted dementia and AD diagnostic criteria
________


THE CASE FOR REVISING THE RESEARCH CRITERIA FOR AD DIAGNOSIS

There are several factors that highlight the need to update
the current research criteria for AD.


Insufficient diagnostic specificity

The DSM-IV-TR and NINCDS–ADRDA criteria have
been validated against neuropathological gold standards
with diagnostic accuracy ranging from 65–96%.[11–14]
However, the specificity of these diagnostic criteria
against other dementias is only 23–88%.[13,14] The accuracy
of these estimates is difficult to assess, given that the
neuropathological standard is not the same in all studies.
Nevertheless, the low specificity must be addressed
through both revised AD and accurate non-AD dementia
diagnostic criteria.


Improved recognition of non-AD dementia

Since the publication of the NINCDS–ADRDA criteria,
operational definition and characterisation of non-AD
dementias has improved. Entities for which there are
diagnostic criteria include the frontotemporal lobar
degenerations (frontotemporal dementia frontal variant,
semantic dementia, progressive non-fluent aphasia),[15–17]
corticobasal degeneration,[18,19] posterior cortical atrophy,[20]
dementia with Lewy bodies,[21] and vascular dementia.[22,23]
Varma and colleagues[13] showed that many of these
disorders can fulfil the NINCDS–ADRDA criteria and it
is likely that they have been included in AD research
studies. Meanwhile, for each of these disorders, criteria
have been developed that aim for high specificity.
The development of disease-specific criteria that are
applicable in some cases before dementia is fully
manifested has enabled the criteria to be used without
going through the two-step process of dementia
recognition (the syndrome) followed by the specific
disease (the aetiology). For example, the identification of
a dementia syndrome is not required for the diagnosis of
primary progressive aphasia, corticobasal degeneration,
or posterior cortical atrophy even though a dementia as
currently defined will occur during or at the end of the
course of these diseases.[3] The histopathological diagnosis
of the non-AD dementias has also advanced. In the
example of frontotemporal lobar degeneration, the
identification of ubiquitin-immunoreactive cytoplasmic
and intranuclear inclusions as an important pathology
in patients has reduced the neuropathological diagnostic
prevalence of dementia lacking distinctive histopathology
from 40% to 10% in autopsy series.[24–26] There is no
doubt that progress in the clinical definition of non-AD
dementia improves the sensitivity of the currently
accepted diagnostic criteria for AD by reducing the level
of uncertainty.


Improved identification of AD phenotype

When the NINCDS–ADRDA criteria were first published,
the authors noted that they were not yet fully operational
because of insufficient knowledge about the disease.[2]
Since then, the clinical phenotype of AD has been much
more clearly elucidated. In most patients with AD
(86–94%), there is a progressive amnestic core that
appears as an impairment of episodic memory.[27–29] The
pathological pathway of Alzheimer’s related changes has
been fully described[30,31] and involves the medial temporal
structures (eg, entorhinal cortex, hippocampal formation,
parahippocampal gyrus) early in the course of the disease.
Moreover, the episodic memory disorder of AD correlates
well with the distribution of neurofibrillary tangles
within the MTL[32] and with MRI volumetric loss of the
hippocampus,[33] structures known to be critical for
episodic memory. The availability of neuroimaging
techniques that can reliably measure the MTL have
further supported this vital cliniconeuroanatomic
correlation.


Need to test early intervention

The rapid growth of knowledge about the potential
pathogenic mechanisms of AD including the
amyloidopathy and tauopathy has spawned numerous
experimental therapeutic approaches to enter into
clinical trials. There is accruing evidence that, years
before the onset of clinical symptoms, there is an AD
process evolving along a predictable pattern of
progression in the brain.[30,31] The neurobiological
advantage of earlier intervention within this cascade is
clear. Earlier intervention with disease-modifying
therapies[34] is likely to be more effective when there is a
lower burden of amyloid and hyperphosphorylated tau
and may truncate the ill effects of secondary events due
to inflammatory, oxidation, excitotoxicity, and apoptosis.
Early intervention may also be directly targeted against
these events because they may play an important part in
early phases of AD. By the time there is clear functional
disability, the disease process is significantly advanced
and even definitive interventions are likely to be
suboptimal. Revised research criteria would allow
diagnosis when symptoms first appear, before full-blown
dementia, thus supporting earlier intervention at the
prodromal stage (figure).


________
Figure: Alzheimer’s disease starts and should be identified before the occurrence of full-blown dementia (as
for other dementing conditions)
AD=Alzheimer’s disease; VD=vascular dementia; FTD=frontotemporal dementia; PPA=primary progressive
aphasia; DLB=dementia with Lewy bodies.
________



Problems with definition of mild cognitive impairment

One of the proposed advantages of mild cognitive
impairment has been its potential usefulness for
clinical trials directed at delaying the time to onset of
AD. A series of large randomised controlled trials with
both non-steroidal anti-inflammatory drugs and acetyl-
cholinesterase inhibitors have sought to establish the
usefulness of these drugs in delaying the conversion of
mild cognitive impairment to AD. However, the lessons
learned have highlighted the problems of mild cognitive
impairment within this type of randomised controlled
trial. With only small variations in the inclusion criteria
for mild cognitive impairment, four trials (ADCS-MIS,
InDDEx, Gal-Int 11, and Rofecoxib) have had a very wide
range of annual rates of progression to AD dementia
(panel 1, 4·8–16%).[35–38] The intention in these trials on
mild cognitive impairment was to include many
individuals with prodromal AD (ie, individuals with
symptoms not sufficiently severe to meet currently
accepted diagnostic criteria) that later progress to meet
these criteria. When the mild cognitive impairment
inclusion criteria of these trials were applied to a cohort
of memory clinic patients in an observational study, they
had diagnostic sensitivities of 46–88% and specificities
of 37–90% in identifying prodromal AD.[39] Given these
numbers, these trials have clearly treated many patients
who do not have AD or are not going to progress to AD
for a long time. This has diluted the potential for a
significant treatment effect and may have contributed to
the negative outcomes where none of these drugs were
successful at delaying the time to diagnosis of AD.[37–38]
These trials have also incurred significant costs where
sample sizes of 750 to 1000 have been called for with
durations of 3–4 years. Increasing the severity of mild
cognitive impairment needed for inclusion in trials
might improve sensitivity, specificity, and predictive
values. However, participants would then be much
closer to the current dementia threshold and would have
a greater pathological burden, making the clinical gain
marginal and disease modification difficult.

Neuropathological findings in mild cognitive
impairment have also reinforced the heterogeneity of the
clinical disorders subsumed under the definition mild
cognitive impairment. To address the recognised clinical
and pathological heterogeneity, it has been proposed that
subtyping of mild cognitive impairment might be
useful.[40,41] The term amnestic mild cognitive impairment
(panel 1) has been proposed to include individuals with
subjective memory symptoms, objective memory
impairment, and with other cognitive domains and
activities of daily living generally assessed as being
normal.[9] However, only 70% of a selected cohort of
people with amnestic mild cognitive impairment
clinically identified to have progressed to dementia
actually met neuropathological criteria for AD.[42] This
finding indicates that applying the criteria for this subtype
of mild cognitive impairment clinically, without other
objective evidence such as neuroimaging or results of
cerebrospinal fluid analyses, will lack specificity for
predicting the future development of AD since at least
30% of these will have non-AD pathology. If 30% of cases
enrolled in a study that assesses drugs targeting amyloid
or neurofibrillary tangles were to have non-AD pathology,
there would be a substantial loss of power and possibly a
conclusion that a medication was not effective. If all cases
could be ascertained as having AD, a positive outcome
might result. Thus, the most accurate determination that
an individual has prodromal AD is critical.

In the planning of trials of disease-modifying treatment,
special care will be needed to limit not only the exposure
of potentially toxic therapies to those with prodromal
AD but also to reliably exclude those who are destined
to develop non-AD dementia. Our proposal for multi-
dimensionally established identification of AD would
have potential superiority to the intrinsically hetero-
geneous state of mild cognitive impairment and would
advance the concept of mild cognitive impairment to its
natural next level of more desirably identifying
prodromal AD.


Unclear distinction between mild cognitive impairment and AD

The transition from mild cognitive impairment to AD has
been an a priori primary endpoint in several randomised
controlled trials.[37,38,43] There is an inherent arbitrariness in
determining a binary outcome, that is, conversion or no
conversion, when the underlying disease is a continuous
process. Individual clinicians’ experience in dementia
diagnosis and the quality of the information they receive
on the cognitive and functional status of patients will
affect the threshold of detection of the transition to AD.[27]
Our revised research criteria will eliminate the mild
cognitive impairment construct, thus bypassing the
binary outcome in the clinical categorisation process
associated with it as well as problems with reliability.


New biomarkers for AD

Over the past two decades since the NINCDS–ADRDA
criteria were published, great progress has been made in
identifying the AD-associated structural and molecular
changes in the brain and their biochemical footprints.
MRI enables detailed visualisation of MTL structures
implicated in the core diagnostic feature of AD.[44] PET
with fluorodeoxyglucose (FDG) has been approved in the
USA for diagnostic purposes and is sensitive and specific
in detecting AD in its early stages.[45] Cerebrospinal fluid
biomarkers for detecting the key molecular pathological
features of AD in vivo are available and can be assessed
reliably.[46] Their diagnostic predictability has been extended
to mild cognitive impairment.[47] In vivo imaging of
pathology-specific proteins (Pittsburgh compound B
[PiB], FDDNP)[48,49] are advancing in their development and
potentially in our ability to accurately identify prodromal
and even preclinical AD (panel 1). The growing body of
evidence about AD biomarkers allows us to incorporate
these into our new diagnostic research criteria for AD.


OBJECTIVES

An international working group was convened in 2005
to discuss the opportunity for developing a diagnostic
framework for AD that would include the prodromal
stages and the integration of biomarkers and to
define the future goals and steps for the validation of
such a framework. This paper provides the consensus
recommendations of the working group and sets out the
framework for revised research criteria for AD that would
apply both in the early stages and across the full spectrum
of the illness.


METHODS

15 international dementia experts were invited by two of
the authors (Dubois and Scheltens) to attend a satellite
workshop at the Second Congress of the International
Society for Vascular Behavioural and Cognitive Disorders
(Vas-Cog) in Florence on June 9, 2005. The participants
were each asked to present the evidence base of published
literature around a range of topics including clinical,
functional, neuropsychiatric and behavioural, cognitive,
neuroimaging, neuropathology, and laboratory markers
pertinent to the early stages of AD, and to provide
expert opinion where there was no published evidence. A
draft document outlining revised diagnostic criteria was
subsequently developed by the lead authors (Dubois,
Feldman, and Scheltens) and then refined in further
correspondence with conference participants. Additional
members were then recruited into the working group
to broaden the perspective before the finalisation and
submission of the current proposed AD research criteria
for publication.


PROPOSED DIAGNOSTIC CRITERIA FOR PROBABLE AD

The proposed framework for revised criteria for probable
AD retains the designation of probable AD. It does not
include a designation of possible AD because of the
incompatibility of this definition with diagnostic criteria
that are highly specific for AD. The framework addresses
the disease presentation that is typical for AD. We exclude
atypical presentations including focal cortical syndromes
(primary progressive aphasia, visuospatial dysfunction)
where an antemortem diagnosis would at best receive
the designation of possible AD from the framework.
This may change in the future as work on diagnostic
biomarkers advances and reliance on a well characterised
clinical phenotype is lessened. In the absence of
completely specific biomarkers, the clinical diagnosis of
AD can still be only probabilistic, even in the case of
typical AD. To meet criteria for probable AD, an affected
individual must fulfil criterion A (the core clinical
criterion) and at least one or more of the supportive
biomarker criteria: B, C, D, or E (panel 2).


________
Panel 2: Diagnostic criteria for AD

Probable AD: A plus one or more supportive features B, C, D, or E

- Core diagnostic criteria

A. Presence of an early and significant episodic memory impairment that includes the following features:
A.1. Gradual and progressive change in memory function reported by patients or informants over more than 6 months
A.2. Objective evidence of significantly impaired episodic memory on testing: this generally consists of recall deficit that does not improve significantly or does not normalise with cueing or recognition testing and after effective encoding of information has been previously controlled
A.3. The episodic memory impairment can be isolated or associated with other cognitive changes at the onset of AD or as AD advances


- Supportive features

B. Presence of medial temporal lobe atrophy
• Volume loss of hippocampi, entorhinal cortex, amygdala evidenced on MRI with qualitative ratings using visual scoring (referenced to well characterised population with age norms) or quantitative volumetry of regions of interest (referenced to well characterised population with age norms)

C. Abnormal cerebrospinal fluid biomarker
• Low amyloid β1–42 concentrations, increased total tau concentrations, or increased phospho-tau concentrations, or combinations of the three
• Other well validated markers to be discovered in the future

D. Specific pattern on functional neuroimaging with PET
• Reduced glucose metabolism in bilateral temporal parietal regions
• Other well validated ligands, including those that foreseeably will emerge such as Pittsburg compound B or FDDNP

E. Proven AD autosomal dominant mutation within the immediate family


- Exclusion criteria

History
• Sudden onset
• Early occurrence of the following symptoms: gait disturbances, seizures, behavioural changes

Clinical features
• Focal neurological features including hemiparesis, sensory loss, visual field deficits
• Early extrapyramidal signs

Other medical disorders severe enough to account for memory and related symptoms
• Non-AD dementia
• Major depression
• Cerebrovascular disease
• Toxic and metabolic abnormalities, all of which may require specific investigations
• MRI FLAIR or T2 signal abnormalities in the medial temporal lobe that are consistent with infectious or vascular insults


- Criteria for definite AD

AD is considered definite if the following are present:
• Both clinical and histopathological (brain biopsy or autopsy) evidence of the disease, as required by the NIA-Reagan criteria for the post-mortem diagnosis of AD; criteria must both be present[139]
• Both clinical and genetic evidence (mutation on chromosome 1, 14, or 21) of AD; criteria must both be present
________


Core diagnostic criterion: early episodic memory impairment (A)

1. Gradual and progressive change in memory function at
disease onset reported by patients or informants for aperiod
greater than 6 months

The reporting of subjective memory complaints is a
common symptom in an ageing population,[50] at a
prevalence that far exceeds the risk of being classified as
having AD. Subjective memory complaints in elderly
people may result from normal ageing or various medical
disorders, and they are commonly associated with
depression.[51–53] However, such self-reported symptoms are
associated with a high risk of future development of AD[54,55]
and, therefore, should be carefully taken into account. The
perceptions of patients’ symptoms from an informant
or proxy are perhaps more significant as they are more
strongly related to objective memory performance[51] and
are predictive of progression to AD.[52] To satisfy criterion A,
memory symptoms must start gradually and show
progressive decline over at least 6 months. Particular
attention should be paid to intraindividual decline, which
improves the identification of those individuals with
prodromal AD.[56]

2. Objective evidence of significantly impaired episodic memory
on testing

A diagnosis of AD requires an objective deficit on
memory testing (recall deficit with intrusions). Tests of
delayed recall discriminate very mild AD from normal
healthy controls with high accuracy (>90%).[57] Such tests
also predict prodromal AD better than other memory or
non-memory measures, with accuracy greater than
80%.[57–63] Delayed recall is a reliable predictor of AD in
individuals with mild cognitive impairment.[64,65] A meta-
analysis of 47 studies calculated pooled effect sizes
between incident AD and control groups. Episodic
memory yielded the largest pooled effect sizes (>1 SD),
along with executive functioning and perceptual speed.[66]
Within episodic memory, delayed recall testing showed a
larger effect size than immediate recall testing.

Impaired delayed recall is not itself evidence of an AD-
related memory disorder. Genuine deficits in encoding
and storage processes that are characteristic for AD must
be distinguished from non-AD deficits that can also
affect delayed recall, including attentional difficulties that
may be present in depression,[67] or inefficient retrieval
strategies associated with normal ageing,[68] frontotemporal
dementia,[69] or subcortical-frontal dementias.[70] The
accurate diagnosis of the episodic memory deficit of AD
can be improved by use of test paradigms that provide
encoding specificity.[71] Within such paradigms, test
materials are encoded along with specific cues—for
example, semantic cues, which are used to control for an
effective encoding and are subsequently presented to
maximise retrieval. Coordinated encoding and retrieval
paradigms of this type include the free and cued
recall test[71] or similar cued recall paradigms.[72,73] Within
these neuropsychological test paradigms, measures
of sensitivity to semantic cueing can successfully
differentiate patients with AD from healthy controls,
even when patients are equated to controls on mini-
mental state examination scores or when disease severity
is very mild.[72,74,75] Buschke and co-workers derived
sensitivity and specificity estimates of 93% and 99%,
respectively, for the discrimination of patients with
mild AD from healthy people with such a strategy.[72]
Patients with non-AD disorders including progressive
supranuclear palsy and Parkinson’s and Huntington’s
diseases do almost as well as control individuals under
encoding specificity conditions whereas patients with
AD do not normalise their recall deficit.[70] Patients with
very mild AD also have a measurable reduction in
sensitivity to cueing.[75] Reduced benefit from cueing at
recall reliably identifies prodromal AD.[73]

3.The episodic memory impairment can be isolated or associated
with other cognitive changes at onset of AD or as AD advances

In most cases, even at the earliest stages of the disease,
the memory disorder is associated with other cognitive
changes. As AD advances, these changes become notable
and can involve the following domains: executive function
(conceptualisation with impaired abstract thinking;
working memory with decreased digit span or mental
ordering; activation of mental set with decreased verbal
fluencies); language (naming difficulties and impaired
comprehension); praxis (impaired imitation, production,
or recognition of gestures); and complex visual processing
and gnosis (impaired recognition of objects or faces).
The emergence of neuropsychiatric symptoms, including
apathy or delusions, also constitutes a clinical marker of
disease.[76,77] Neuropsychiatric symptoms cannot be a core
diagnostic feature because they are less specific and
generally occur at a high prevalence later in the course of
the disease. When there is evidence of impairment in
multiple cognitive domains, functional disability, and
neuropsychiatric symptoms, a more widespread diffusion
of neuronal lesions in cortical and subcortical structures
can be established.[30] However, even in these more
advanced cases, there should be evidence of an early
and previous episodic memory deficit as a mandatory
requirement for the diagnosis of AD.


Supportive features

- Atrophy of medial temporal structures on MRI (B)

Atrophy of the MTL on MRI seems to be common in AD
(71–96%, depending on disease severity), frequent in
mild cognitive impairment (59–78%), but less frequent
in normal ageing (29%).[78,79] MTL atrophy is related to the
presence of AD neuropathological and its severity, both
in terms of fulfilment of AD neuropathological criteria
and Braak stages.[44,80] MRI measurements of MTL
structures include qualitative ratings of the atrophy in
the hippocampal formation[81] or quantitative techniques
with tissue segmentation and digital computation of
volume.[82] Both techniques can reliably separate AD group
data from normal age-matched control group data, with
sensitivities and specificities greater than 85%.[83–85]
Hippocampal volumes can distinguish AD equally at
younger (≤70 years) and old ages (>70 years).[86] Qualitative
measures have been useful in distinguishing patients
with AD from those with non-AD dementia including
vascular, frontotemporal, and dementia of unspecified
cause, with combined mini-mental state examination
and MTL atrophy ratings yielding sensitivity and
specificity greater than 85%.[87]

In studies of mild cognitive impairment, the accuracy
of MTL atrophy measures in identifying prodromal AD
has been generally lower, possibly because individuals
who did not meet currently accepted AD diagnostic
criteria at study completion included some cases that
would have done so at a later time. Qualitative MTL
ratings can identify prodromal AD; however the
sensitivities and specificities, respectively, of 51–70% and
68–69%[88–91] at present limit their usefulness. The
predictive usefulness of quantitative measures of
hippocampal volume in identifying prodromal AD is
inconsistent.[88,89,92–94] Measures of hippocampal subfields
might be more useful than measures of the entire
structure.[95,96] Other structures or combinations of
structures within and beyond the MTL may prove to be
more sensitive to early AD pathology. For example,
entorhinal cortex volume identifies prodromal AD more
accurately than hippocampal volume, with a sensitivity
of 83% and specificity of 73%.[97] There are, however,
technical difficulties in measuring this region that must
be resolved.[98,99] Combinations of MTL volumes and lateral
temporal lobe or anterior cingulate volumes also detect
prodromal AD with variable success (sensitivity 68–93%,
specificity 48–96%).[100,101]

There is a strong correlation between MTL volumes
and episodic memory performance.[33,102] In turn, there is a
potential incremental value of MTL measurement beyond
the episodic memory impairment in the identification
of prodromal AD. In several studies, MTL measures
(quantitative and qualitative) contributed independently
of memory scores to the identification of prodromal
AD.[90,93] The reported accuracy of identifying prodromal
AD increased from 74% to 81%[89] and from 88% to 96%[88]
when MTL measures were added to age and memory
scores, respectively.

Inclusion of MTL atrophy as a diagnostic criterion of
AD, irrespective of the age at onset,[86] mandates exclusion
of other causes of MTL structural abnormality including
bilateral ischaemia, bilateral hippocampal sclerosis,
herpes simplex encephalitis, and temporal lobe epilepsy.
T2 weighted MRI, coupled to history and examination,
and potentially adjunctive directed tests such as cerebro-
spinal fluid analysis and EEG should facilitate this
discrimination.[103,104]

- Abnormal cerebrospinal fluidbiomarkers (C)

In the NINCDS–ADRDA guidelines, cerebrospinal fluid
examination was recommended as an exclusion
procedure for non-AD dementia, due to inflammatory
disease, vasculitis, or demyelination.[2] Since then, there
has been a lot of research into the usefulness of
AD-specific biomarkers that are reflective of the
central pathogenic processes of amyloid β aggregation
and hyperphosphorylation of tau protein. These markers
have included amyloid β1–42 (Aβ42), total tau (t-tau), and
phospho-tau (p-tau).[105–107] In AD, the concentration of
Aβ42 in cerebrospinal fluid is low and that of t-tau is
high compared with those in healthy controls.[105,106]
Concentrations of different phosphorylated tau epitopes
may also be high.[108,109] Aβ42 concentration in the
cerebrospinal fluid is normal in patients with depression
and decreased in dementia with Lewy bodies, fronto-
temporal lobar degeneration, and vascular dementia.[110]
This lack of specificity is not fully explained, but may
relate to the lack of histopathological verification or the
presence of comorbid AD. T-tau concentration is normal
in depression, may be slightly raised in dementia with
Lewy bodies and frontotemporal lobal degeneration, and
is very high in Creutzfeldt-Jakob disease.[110,111] Measurement
of the concentration of p-tau, notably p-tau 231, increases
the specificity for AD, especially in contrast to fronto-
temporal lobar degeneration.[109] The pooled sensitivity
and specificity for Aβ42 in AD versus controls from
13 studies involving 600 patients and 450 controls were
86% and 90%, respectively.[110] For t-tau, the sensitivity was
81% and the specificity 90%, pooled from 36 studies with
2500 patients and 1400 controls.[110] Across 11 studies with
a total of 800 patients and 370 controls, p-tau had a mean
sensitivity of 80% when specificity was set at 92%, but
sensitivities varied widely among studies using different
methods.[110] By use of a combination of concentrations of
Aβ42 and t-tau for AD versus controls, high sensitivities
(85–94%) and specificities (83–100%) can be reached.[110]

Several recent studies have specifically addressed the
value of cerebrospinal fluid biomarkers in identifying
prodromal AD. Combinations of abnormal markers (low
Aβ42, high t-tau, high p-tau 181) reached a hazard ratio of
17 to 20 for predicting AD in a follow-up of 4–6 years.[47]
Sensitivities and specificities in this study were >90%
and >85%, respectively, which agreed with those in a
similar one with much shorter follow-up (1 year).[47,112] This
high diagnostic usefulness of cerebrospinal fluid markers
in the mild cognitive impairment stage supports their
incremental value over memory impairment in the
diagnostic scheme and justifies their inclusion as a
diagnostic criterion.

Using an adapted spinal needle (Sprotte 24 g), lumbar
puncture can be done with a very low rate of clinically
significant adverse events and with a good acceptability
in cognitively impaired people and healthy adults of all
ages.[113]

- Specific metabolic pattern evidenced with molecular neuroimaging methods (D)

PET and single photon emission computed tomography
(SPECT) are in vivo nuclear radioisotopic scans that can
measure blood flow (99mTc-HMPAO or 133Xe), glucose
metabolism (18F-FDG PET), and, more recently, protein
aggregates of amyloid and tau. Within an AD diagnostic
framework their ideal role is to increase the specificity of
clinical criteria.

A reduction of glucose metabolism as seen on PET in
bilateral temporal parietal regions and in the posterior
cingulate is the most commonly described diagnostic
criterion for AD.[114] A recent meta-analysis of nine studies
reported that for the discrimination of patients with AD
from healthy controls, pooled sensitivities and specificities
were 86% for temporoparietal hypometabolism. There is
a wide range for both sensitivities and specificities,
without clear explanation of this variability.[115] When
histopathological examination has been used as the
gold standard, sensitivity is 88–95% and specificity is
62–74%.[114,115] Because of cognitive reserve, the reduction
in temporoparietal glucose metabolism diagnostic for
AD might vary with educational attainment or IQ at any
level of clinical severity.[116]

PET has been successful in distinguishing dementia
with Lewy bodies from AD, with sensitivity of 86–92%
and specificity of 80–81% when visual-association cortex
was considered.[117,118] Discrimination from frontotemporal
dementia has also been achieved, with sensitivity and
specificity of 78% and 71%.[119] There has been limited
accuracy in differentiating AD from vascular dementia
(sensitivity 75–88%, specificity 18–53%).[120,121]

The usefulness of FDG-PET in the detection of
prodromal AD has only just begun to be addressed in
studies with small samples of patients with mild cognitive
impairment and limited follow-up (≤3 years). Metabolic
reductions in the anterior cingulate, posterior cingulate,
and temporal, parietal, and medial temporal cortices
detected prodromal AD, with accuracy estimates ranging
from 75% to 84%.[122–124] The potential incremental value
of PET over other diagnostic markers in identifying
prodromal AD is poorly defined. PET may be more
accurate when delayed recall scores are severely impaired
(sensitivity and specificity >90%).[125]

There are promising PET techniques that provide
in-vivo visualisation of amyloid and potentially
neurofibrillary tangles. Studies using PiB (N-methyl-
[11C]2-(4-methylaminophenyl)-6-hydroxybenzothiazole)
and FDDNP (2-(1-[6-[(2-[18F]fluoroethyl](methyl)amino]-2-
naphthyl]ethylidene)malononitrile) have shown a pattern
of increased radioligand retention in patients with AD
compared with control individuals that is consistent with
AD pathology.[48,49,126] Furthermore, positive cortical PiB
binding has been associated with low cerebrospinal fluid
Aβ42 concentrations in AD.[127] These protein visualisation
techniques clearly have the potential of increasing the
usefulness of PET in AD within the diagnostic framework,
but their diagnostic accuracy, in particular their specificity
for AD, requires further investigation as there is evidence
of high AD-like PiB retention in some healthy people
and some people with mild cognitive impairment.[127–129]
AD-like PiB retention in healthy people might signal a
preclinical AD state in asymptomatic individuals who
later meet currently accepted dementia and AD diagnostic
criteria, whereas in mild cognitive impairment it might
reveal prodromal AD. Longitudinal follow-up is essential
for the verification of the presumption that these are
indeed preclinical and prodromal AD cases and not false
positives.[128]

Because SPECT is more widely available and cheaper
than PET, it has received much attention as an alternative
to PET. However, at present, the technique is not included
in these proposed criteria as the diagnostic accuracy
estimates for this modality generally fall below the
requisite 80% levels specified by the Reagan Biomarker
Working Group.[130] 99M Tc-HMPAO SPECT identifies
diagnosed AD with moderate sensitivity (77–80%) and
specificity (65–93%). A pattern of bilateral temporal
parietal hypoperfusion increases diagnostic certainty
over clinical diagnosis alone.[131]

According to a recent meta-analysis SPECT
distinguished AD from non-AD in studies including
healthy controls with pooled weighted sensitivities
ranging from 65% to 71%, with a specificity of 79%.[132]
There are few SPECT studies that have adequately
addressed the comparison between AD and non-AD
dementias. The few that did provided a pooled weighted
sensitivity and specificity for AD versus frontotemporal
dementia of 71% and 78%, respectively, and for AD
versus vascular dementia of 71% and 75%, respectively.[132]
More specific ligand methods, for example dopamine
SPECT scanning with fluoropropyl-CIT, may have
particular utility in distinguishing dementia with Lewy
bodies and Parkinson’s disease dementia from AD
(sensitivity 88%, specificity 85%).[133–135]

Two small retrospective SPECT studies of mild
cognitive impairment suggest that hypoperfusion in
parietal and temporal lobe regions, and in the precuneus,
may be brain functional patterns occurring very early in
AD. In both studies, patterns of regional blood flow on
SPECT distinguished prodromal AD with accuracy
greater than 80%.[135,136] These studies require replication
with larger samples and prospective methodology before
the technique can become a recommended criterion. The
thioflavin derivative IMPY (6-iodo-2-(4-dimethylamino-)
phenyl-imidazo[1,2-a]pyridine), which targets amyloid
plaques for in vivo imaging in SPECT has not yet been
investigated in living patients with either AD or mild
cognitive impairment, but this ligand might help
measure amyloid plaque burden in the future, thus
providing an additional new approach and usefulness for
functional imaging.[137]

- Familial genetic mutations (E)

Three autosomal dominant mutations that cause AD
have been identified on chromosomes 21 (amyloid
precursor protein), 14 (presenilin 1), and 1 (presenilin
2).[138] The presence of a proband with genetic-testing
evidence of one of these mutations can be considered as
strongly supportive for the diagnosis of AD for affected
individuals within the immediate family who did not
themselves have a genetic test for this mutation. If
individuals with a positive mutation history of the
described type present with the core amnestic criterion
A, they will be considered to meet criteria for AD within
the revised diagnostic framework. The gold standard for
definite diagnosis of AD in this setting would of course
be genetic testing and verification of a genetic mutation
in these individuals.


Exclusion criteria

Probable AD diagnosis cannot be established if the illness
begins with a sudden onset, has focal neurological
findings including hemiparesis, sensory loss, visual field
deficits, or where there are seizures, gait disturbances, or
extrapyramidal signs at the onset or very early in the
course of the illness. Other medical, neurological, or
psychiatric disorders that could otherwise account for the
deterioration in memory and related symptoms must
be excluded. The diagnosis should be questioned in
case of the following red flags: early behavioural
disturbances (particularly disinhibition, euphoria, or
psychosis), early extrapyramidal symptoms, early visual
hallucinations, early visuospatial impairment, marked
fluctuations in cognition and REM sleep behavioural
disorders. The presence of cerebrovascular lesions,
particularly white-matter lesions and lacunar infarctions
both symptomatic and asymptomatic, are common with
ageing. To establish probable AD, cerebrovascular disease
that is sufficiently severe to account for the cognitive and
functional deficits must be excluded. Probable AD cannot
be diagnosed if the symptom profile suggestive of
dementia with Lewy bodies (pronounced fluctuations in
attention and cognition, recurrent prominent visual
hallucinations, and motor parkinsonism) or if any other
non-AD dementia is present. The presence of a delirium
or toxic metabolic cause for the cognitive disorder
precludes a diagnosis of probable AD (at least until the
delirium has cleared) as does an unexplained altered state
of consciousness.


DISCUSSION

This working group has identified, by consensus, that
new research criteria are timely, realistic, and feasible.
Our proposed AD diagnostic framework (panel 2)[139] is
anchored around a core clinical phenotype supported
by brain-structure abnormalities, molecular imaging
impairment, biochemical changes, or genetic mutations
associated with AD. The timeliness of these criteria is
underscored by the many drugs in development that are
directed at changing the disease pathogenesis through
amyloid immunotherapy, gamma or beta secretase
inhibitors and modulators, alpha secretase activators, tau
kinase inhibitors, and nerve growth factors. Further new
approaches directed at tau and tangles are foreseen.
There is a neurobiological imperative to identify
AD before the point of disease where irreversible
pathological injury would prevent effective intervention.[140]
The proposed criteria should allow an earlier and more
specific AD diagnosis than their predecessor, the
NINCDS–ADRDA criteria.

These proposed criteria move away from the traditional
two-step approach of first identifying dementia according
to degree of functional disability, and then specifying its
cause. Rather, they aim to define the clinical, biochemical,
structural, and metabolic presence of AD. The cornerstone
clinical criterion A specifies that there is an episodic
memory deficit within test conditions of encoding
specificity. This criterion should allow 86–94% of cases to
be included.[28,29] Beyond this core criterion, the presence
of at least one biological footprint of the disease, either by
criterion B (structural imaging), criterion C (cerebrospinal
fluid), criterion D (molecular imaging), or criterion E
(dominant mutation within the immediate family) is also
needed to establish a positive diagnosis. The requirement,
for diagnosis, of a clinical phenotype in combination
with any one of the supportive features currently
represents the most balanced approach because the
clinical phenotype of AD is better known than its
biological phenotype. We have no empirical basis at this
time for assigning different weightings to the supportive
features or recommending combinations of features or,
alternatively, requiring the presence of all. However, as
new evidence accrues on biological markers for AD,
especially those detecting AD-pathology specific markers
such as amyloid imaging, the weighting of the supportive
features may change. Other combinations may prove to
have greater diagnostic accuracy or new features may be
introduced. This will evolve as data sets gathered with all
modalities are assessed.

We recognise that these criteria represent a cultural
shift requiring more biologically focused work-up than
previous approaches; however, this seems to be the best
way to integrate the profound advances into the clinical
arena. When effective disease-modifying medications
are available, the argument for such biologically based
studies will be even more compelling. Some research
needs will be better addressed with a more stringent
approach requiring that each diagnostic criterion be met.
For example, proof-of-principle studies may benefit from
the most highly selected AD study samples where the
presence of all supportive features might be specified.
This could maximise specificity for AD, but impose a
substantial loss of sensitivity that would need to be
re-addressed in later stages of development.

There are important differential diagnostic challenges
that can be anticipated with the application of the
proposed criteria. Of primary concern are non-AD
amnestic disorders that can be associated with MTL
damage including bilateral ischaemic injury, hippocampal
sclerosis, limbic encephalitis, and temporal lobe epilepsy.
Non-AD neurodegenerative disorders including tangle-
only pathological changes and argyrophilic grain disease
may also involve the MTL and limbic system.[42] Depression
can present with episodic memory impairment and MRI
changes in hippocampal volume.[141] These disorders could
potentially satisfy criteria A and B, and in turn, where
possible, must be ruled out in each instance. Careful
clinical assessment with the use of specific memory tests,
careful attention to T2 signal abnormalities within the
MTL, and other investigations as clinically indicated will
be called for to establish the AD diagnosis. We also
cannot exclude that the non-AD dementias including
frontotemporal lobar degeneration, vascular dementia,
and dementia with Lewy bodies may in some cases have
the core clinical amnestic presentation specified in this
framework. These non-AD dementias may also have
positive molecular imaging or cerebrospinal fluid
findings as has been shown in the reviewed evidence. If a
non-AD cause is suspected, it must be ruled out carefully
on a case-by-case basis by applying in parallel the
diagnostic criteria for the other disorders.

We have specifically not addressed the issue of mixed
disease for two reasons. First, in many instances mixed
stands for comorbid disease being present but not
being the principal cause of the dementia syndrome. As
such, nothing has changed over current practice.
Second, by narrowing the definition to a strict memory
presentation with additional evidence for underlying
AD pathological change, the chance of a patient
fulfilling more than one set of criteria, as was the case
with the NINCDS–ADRDA criteria, has actually been
reduced. Even when a patient has abundant white-
matter changes on MRI, thought to be of vascular
origin, the presence of criterion A and the absence of
overt dementia in the sense of the NINDS–AIREN
criteria (ie, involvement of other cognitive domains),
renders the patient more likely to have AD (with
concomitant vascular disease) than vascular dementia.
The MRI changes may still be used to guide therapy
towards secondary prevention.

Moreover, the proposed criteria depict typical AD.
There are also atypical forms of neuropathologically
confirmed AD.[28,29] These forms clearly deviate from
the described amnestic presentation and include focal
cortical syndromes, particularly posterior cortical
degeneration where there is visual or visuospatial
impairment, or frontal forms with prominent behavioural
symptoms. The non-memory clinical phenotype might
be influenced by the apolipoprotein genotype.[142] Estimates
of the relative prevalence of these atypical presentations
have ranged from 6% to 14%.[28,29] These presentations will
still clearly elude diagnosis according to the revised
criteria as they did in the NINCDS–ADRDA criteria.
Their inclusion in research protocols remains too
uncertain to be made with sufficient reliability and for
this reason we have excluded them from the present
diagnostic framework.

The strength of these proposed research criteria is the
introduction of neurobiological measures on to the
clinically based criteria. There are, however, many
limitations and steps still needed. In their current
formulation, these proposed diagnostic criteria still
require decisions around how they are to be put into
practice. For example, for the core criterion of significant
episodic memory impairment, we have identified the
memory test paradigms that can distinguish AD-
associated deficits from other memory difficulties, but
we have not defined a magnitude of deficit or the
comparative norms that should be used. In structural
imaging, we have not presented a specific best test or
method for MTL atrophy. There remains uncertainty as
to the most effective method of assessment, qualitative
or quantitative, and for the latter, the specific region
within the MTL for measurement. There is no
specification of the amount of atrophy that is optimally
diagnostic of AD. Within molecular neuroimaging, there
are similar open questions with regard to which regions
are optimally diagnostic, whether a qualitative versus
quantitative approach should be taken, and what degree
of hypometabolism is diagnostic. Finally, we have
not specified which cerebrospinal fluid marker or
combination of markers should be used to support a
positive diagnosis. Concentrations of cerebrospinal fluid
markers vary substantially with different assays but also
with the same assay done in different centres,[110] raising
important questions about measurements and sources
of error. Although most of these questions will receive
empirical answers in the future, this will not entirely
resolve an issue that is also philosophical, around
whether an approach primarily based on applied clinical
judgment or one based on fully operational definitions
will work better for the research diagnosis of AD. At this
point we favour the former approach of clinical judgment
being applied to the determination of each criterion.

Validation studies of the proposed diagnostic criteria
will clearly be needed because, inherent in this new
definition of AD is the assumption that they indicate the
presence of the neurodegenerative process of AD,
including those cases presenting very early in the course
of the disease. Their validity will need to be established
for different types of discrimination and at different time
points in the disease course, including discrimination of
early AD from normal ageing, and AD from non-AD
dementias. Two major strategies can be used for the
validation of the proposed criteria. First, the criteria can
be applied retrospectively to existing cohorts that have
detailed investigations including neuropsychology, MRI,
cerebrospinal fluid analysis, and PET. The current large
cohort studies of the European Alzheimer Disease
Consortium (EADC), the Alzheimer Disease Cooperative
Study (ADCS), the Alzheimer Disease Neuroimaging
Initiative (ADNI), and other ongoing studies[88,143] will
provide ideal sources within which to validate the
diagnostic criteria of probable AD. Because these are all
multicentre studies, they will also allow the assessment of
measurement reliability across centres as a step towards
standardising measurements and putting our proposed
criteria into practice. Second, new prospective cohorts
should be acquired that are followed to post-mortem. This
prospective validation approach will need to focus on
non-demented individuals with and without cognitive
complaints. At their initial study visit, individuals should
be assessed in parallel with the newly proposed criteria
forADandwiththecriteriaformildcognitiveimpairment.
In subsequent visits, the stability of the diagnosis under
these proposed criteria should be determined. In addition,
the standard NINCDS–ADRDA criteria for AD should be
applied. The first validation measure will be the sensitivity
and specificity of our proposed criteria, obtained at
baseline, for predicting cognitive decline as well as
eventually meeting the existing AD criteria. In addition,
the present criteria should be compared with the
NINCDS–ADRDA criteria to verify whether we achieve
the goal of increasing the specificity for diagnosis.
Ultimately, the sensitivity and specificity for the
pathological diagnosis of AD and other dementias will
need to be assessed. In all future studies using these
criteria, the supportive features used and the number of
patients that have a positive MRI, or cerebrospinal fluid
test, or PET or SPECT should be specified. Validation
studies will necessarily begin with selected samples—for
example those accrued in the ADCS, EADC, and ADNI—
but validation will eventually need to be extended to
unselected heterogeneous community samples. The most
informative studies are those that will use the four criteria
on patients at different stages of the disease, including
the prodromal stage, with a long-term follow-up including
post-mortem examination.

We recognise that the proposed research criteria require
significant expertise, technical skills, and financial
resources to allow the comprehensive assessment of MRI,
PET, and cerebrospinal fluid. MRI will be contraindicated
in some patients or not easily available in some countries,
as may be the case for cerebrospinal fluid biomarkers or
molecular neuroimaging with PET or SPECT. The
multidisciplinary approach that is required for our
diagnostic framework may not yet be feasible in all
memory clinics and certainly not in most epidemiological
studies. The validation studies being proposed will need
to take place within highly specialised AD centres, and if
successful, the research criteria will need to be adapted
for use in standard clinical settings. We foresee that
technically less demanding criteria for clinical settings
might develop from the more technically challenging
research criteria once these are validated.

Finally, these proposed criteria acknowledge the
progress that has been made in the past two decades
in refining our understanding of the neurobiology and
clinical phenomenology of AD. Their usefulness will
be determined in the future as investigators apply the
criteria in a variety of research studies and as key issues
in their application are resolved.


CONTRIBUTORS

Each of the authors has contributed to the writing and the revision of the
paper and has approved the final version. In addition, most of the
authors (16 out of 19) participated in the Florence meeting (June 2005)
and in the related discussions about the revised criteria.


CONFLICTS OF INTEREST

We have no conflicts of interest.

ACKNOWLEDGMENTS

The conference for the international working group held on June 9, 2005,
was sponsored through a grant-in-aid received from Eisai, Janssen-Cilag,
Novartis, and Servier. BD obtained financial support from several
pharmaceutical companies (Servier, Eisai, Novartis, Lundbeck, and
Janssen-Cilag) for the extra costs needed to organise this 1 day satellite
symposium, which included airplane tickets and accommodation for the
following experts: B Dubois, H H Feldman, P Barberger-Gateau,
A Delacourte, S Gauthier, G Jicha, K Meguro, P Robert, M Rossor,
S Salloway, Y Stern, and P Visser. We thank Marsel-Mesulam for helpful
comments on the article.


REFERENCES

1 American Psychiatric Association. Diagnostic and statistical manual
of mental disorders (IV-TR), 4th edn—text revised. Washington,
DC: 2000.

2 McKhann G, Drachman DA, Folstein M, Katzman R, Price DL,
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. Neurology 1984; 34: 939–44.

3 Dubois B, Albert ML. Amnestic MCI or prodromal Alzheimer’s
disease? Lancet Neurol 2004; 3: 246–48.

4 Crook T, Bartus R, Ferris SH, et al. Age-associated memory
impairment: proposed diagnostic criteria and measures of clinical
change—report of a National Institute of Mental Health Work
Group. Dev Neuropsychol 1986; 2: 261–76.

5 Levy R. Aging-Associated Cognitive Decline. Int Psychogeriatr 1994;
6: 63–68.

6 World Health Organization. ICD-10: international statistical
classification of diseases and related health problems—based on
recommendations of the tenth revision conference, 1989 and
adopted by the forty-third World Health Assembly, 10th edn. World
Health Organization; Geneva: 1992.

7 Ebly EM, Hogan DB, Parhad IM. Cognitive impairment in the
non-demented elderly: results from the Canadian study of health
and aging. Arch Neurol 1995; 52: 612–19.

8 Flicker C, Ferris SH, Reisberg B. Mild cognitive impairment in the
elderly: predictors of dementia. Neurology 1991; 41: 1006–09.

9 Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG,
Kokmen E. Mild cognitive impairment: Clinical characterization
and outcome. Arch Neurol 1999; 56: 303–08.

10 Larrieu S, Letenneur L, Orgogozo JM et al. Incidence and outcome
of mild cognitive impairment in a population-based prospective
cohort. Neurology 2002; 59: 1594–99.

11 Lim A, Tsuang D, Kukull W, et al. Clinico-neuropathological
correlation of Alzheimer’s disease in a community-based case
series. J Am Geriatr Soc 1999; 47: 564–69.

12 Petrovitch H, White LR, Ross GW, et al. Accuracy of clinical criteria
for AD in the Honolulu-Asia Aging Study, a population-based study.
Neurology 2001; 57: 226–34.

13 Varma AR, Snowden JS, Lloyd JJ, Talbot PR, Mann DM, Neary D.
Evaluation of the NINCDS–ADRDA criteria in the differentiation of
Alzheimer’s disease and frontotemporal dementia.
J Neurol Neurosurg Psychiatry 1999; 66: 184–88.

14 Kazee AM, Eskin TA, Lapham LW, Gabriel KR, McDaniel KD,
Hamill RW. Clinicopathologic correlates in Alzheimer disease:
assessment of clinical and pathologic diagnostic criteria.
Alzheimer Dis Assoc Disord 1993; 7: 152–64.

15 Neary D, Snowden J, Mann D. Frontotemporal dementia.
Lancet Neurol 2005; 4: 771–80.

16 Hodges JR, Patterson K, Oxbury S, Funnell E. Semantic dementia:
progressive fluent aphasia with temporal lobe atrophy. Brain 1992;
115: 1783–806.

17 Mesulam MM. Slowly progressive aphasia without generalized
dementia. Ann Neurol 1982; 11: 592–98.

18 Rebeiz JJ, Kolodny EH, Richardson EP Jr. Corticodentatonigral
degeneration with neuronal achromasia. Arch Neurol 1968; 18: 20–33.

19 Gibb WR, Luthert PJ, Marsden CD. Corticobasal degeneration.
Brain 1989; 112: 1171–92.

20 Benson DF, Davis RJ, Snyder BD. Posterior cortical atrophy.
Arch Neurol 1988; 45: 789–93.

21 McKeith IG, Galasko D, Kosaka K et al. Consensus guidelines for
the clinical and pathological diagnosis of dementia with Lewy
bodies (DLB): report of the consortium on DLB international
workshop. Neurology 1996; 47: 1113–24.

22 Roman GC, Tatemichi TK, Erkinjuntti T et al. Vascular dementia:
diagnostic criteria for research studies: report of the NINDS-AIREN
international workshop. Neurology 1993; 43: 250–60.

23 Chui HC, Victoroff JI, Margolin D, Jagust WJ, Shankle R,
Katzman R. Criteria for the diagnosis of ischemic vascular
dementia proposed by the State of California Alzheimer’s Disease
Diagnostic and Treatment Centers. Neurology 1992; 42: 473–80.

24 Knopman DS. Overview of dementia lacking distinctive histology:
pathological designation of a progressive dementia. Dementia 1993;
4: 132–36.

25 Forman MS, Farmer J, Johnson JK et al. Frontotemporal dementia:
clinicopathological correlations. Ann Neurol 2006; 596: 952–62.

26 Josephs KA, Holton JL, Rossor MN et al. Frontotemporal lobar
degeneration and ubiquitin immunohistochemistry.
Neuropathol Appl Neurobiol 2004; 30: 369–73.

27 Morris JC. Mild cognitive impairment is early-stage Alzheimer
disease: time to revise diagnostic criteria. Arch Neurol 2006; 63: 15–16.

28 Lopez OL, Becker JT, Klunk W, et al. Research evaluation and
diagnosis of probable Alzheimer’s disease over the last two decades:
I. Neurology 2000; 55: 1854–62.

29 Galton CJ, Patterson K, Xuereb JH, Hodges JR. Atypical and typical
presentations of Alzheimer’s disease: a clinical, neuropsychological,
neuroimaging and pathological study of 13 cases. Brain 2000;
123: 484–98.

30 Braak H, Braak E. Neuropathological staging of Alzheimer-related
changes. Acta Neuropathol (Berl) 1991; 82: 239–59.

31 Delacourte A, David JP, Sergeant N et al. The biochemical pathway
of neurofibrillary degeneration in aging and Alzheimer’s disease.
Neurology 1999; 52: 1158–65.

32 Guillozet AL, Weintraub S, Mash DC, Mesulam MM.
Neurofibrillary tangles, amyloid, and memory in aging and mild
cognitive impairment. Arch Neurol 2003; 60: 729–36.

33 Deweer B, Lehericy S, Pillon B et al. Memory disorders in probable
Alzheimer’s disease: the role of hippocampal atrophy as shown with
MRI. J Neurol Neurosurg Psychiatry 1995; 58: 590–97.

34 Greig NH, Lahiri DK, Giacobini E. Editorial: advances in Alzheimer
therapy: something old, something new, something borrowed,
something blue. Curr Alzheimer Res 2005; 2: 275–79.

35 Jelic V, Kivipelto M, Winblad B. Clinical trials in mild cognitive
impairment: lessons for the future. J Neurol Neurosurg Psychiatry
2006; 77: 429–38.

36 Grundman M, Petersen RC, Bennett DA et al.