doi:10.1136/jnnp.66.2.184 1999;66;184-188 J. Neurol. Neurosurg. Psychiatry

  A R Varma, J S Snowden, J J Lloyd, P R Talbot, D M A Mann and D Neary  

frontotemporal dementiadifferentiation of Alzheimer's disease and Evaluation of the NINCDS-ADRDA criteria in the

http://jnnp.bmj.com/cgi/content/full/66/2/184Updated information and services can be found at:

These include:

References

http://jnnp.bmj.com/cgi/content/full/66/2/184#otherarticles32 online articles that cite this article can be accessed at:  

http://jnnp.bmj.com/cgi/content/full/66/2/184#BIBLThis article cites 21 articles, 14 of which can be accessed free at:

Rapid responses http://jnnp.bmj.com/cgi/eletter-submit/66/2/184

You can respond to this article at:

serviceEmail alerting

top right corner of the article Receive free email alerts when new articles cite this article - sign up in the box at the

Topic collections

(624 articles) Psychiatry of old age � (2023 articles) Memory disorders (psychiatry) �

(4212 articles) Stroke � (311 articles) Neuroimaging �

(3485 articles) Drugs: CNS (not psychiatric) � (1672 articles) Memory disorders (neurology) �

  Articles on similar topics can be found in the following collections

Notes  

http://journals.bmj.com/cgi/reprintformTo order reprints of this article go to:

http://journals.bmj.com/subscriptions/ go to: Journal of Neurology, Neurosurgery, and PsychiatryTo subscribe to

on 9 June 2009 jnnp.bmj.comDownloaded from

Evaluation of the NINCDS-ADRDA criteria inthe diVerentiation of Alzheimer’s disease andfrontotemporal dementia

A R Varma, J S Snowden, J J Lloyd, P R Talbot, D M A Mann, D Neary

AbstractObjectives—The diagnosis of Alzheimer’sdisease (AD) is now reliant on the use ofNINCDS-ADRDA criteria. Other diseasescausing dementia are being increasinglyrecognised—for example, frontotemporaldementia (FTD). Historically, these disor-ders have not been clearly demarcatedfrom AD. This study assesses the capabil-ity of the NINCDS-ADRDA criteria toaccurately distinguish AD from FTD in aseries of pathologically proved cases.Methods—The case records of 56 patients(30 with AD, 26 with FTD) who hadundergone neuropsychological evalua-tion, brain imaging, and ultimately post-mortem, were assessed in terms ofwhether at initial diagnosis the NINCDS-ADRDA criteria were successful in diag-nosing those patients who had AD andexcluding those who did not.Results—(1) The overall sensitivity of theNINCDS-ADRDA criteria in diagnosing“probable” AD from 56 patients with cor-tical dementia (AD and FTD) was 0.93.However, the specificity was only 0.23;most patients with FTD also fulfilledNINCDS-ADRDA criteria for AD. (2)Cognitive deficits in the realms of orienta-tion and praxis significantly increased theodds of a patient having AD comparedwith FTD, whereas deficits in problemsolving significantly decreased the odds.Neuropsychological impairments in thedomains of attention, language, percep-tion, and memory as defined in theNINCDS-ADRDA statement did not con-tribute to the clinical diVerentiation of ADand FTD.Conclusion—NINCDS-ADRDA criteriafail accurately to diVerentiate AD fromFTD. Suggestions to improve the diagnos-tic specificity of the current criteria aremade.(J Neurol Neurosurg Psychiatry 1999;66:184–188)

Keywords: Alzheimer’s disease; frontotemporal demen-tia; NINCDS-ADRDA criteria

Accurate clinical diagnosis of the dementias iscritical for proper management and assessmentof prognosis. With recent advances in therapyand availability of new drugs (for example,donepezil) this exercise is no longer merelyacademic. The two principal causes of “corti-cal” dementias are Alzheimer’s disease (AD)and frontotemporal dementia (FTD). FTD is

characterised by profound alteration in person-ality and social conduct and cognitive impair-ments predominantly in frontal lobe “execu-tive” functions.1 2 By contrast, in ADprominent impairments in the domains ofmemory, visuospatial functions, and languagetypically occur in the context of well preservedsocial skills.3 4 Despite striking diVerences intheir neuropsychological profile there is evi-dence that clinical diVerentiation remainspoor. For example in a study of 21 patientswith pathological confirmation of Pick’s dis-ease (one of the causes of FTD), 85% weremisdiagnosed during life, their condition mis-attributed to AD.5 In another pathologicalstudy of 170 patients with a clinical diagnosisof AD, 12% of patients were misdiagnosedduring life, of whom 19% were found to haveFTD.6

The NINCDS-ADRDA criteria7 have beenwidely adopted and relied upon for the diagno-sis of AD. It is not known whether the use ofthese criteria alone would eVectively excludeFTD. The purpose of the study was todetermine the ability of the criteria to specify atinitial presentation “probable” AD and toexclude non-AD (FTD) from within a group ofthose two conditions.

MethodsThe case records of 56 consecutive patientswith pathological confirmation of either AD orFTD were analysed. All of these patients hadbeen assessed and reviewed systematically dur-ing life in the Cerebral Function Unit, Neurol-ogy Department at the Manchester Royal Infir-mary as part of a prospective longitudinal studyof dementia. The findings of neuropsychologi-cal assessments8 performed at initial patientpresentation were documented by a neurologist(ARV) who was unaware either of the diagnosisor the patient. Information was carefullyrecorded on standardised forms with respect to(a) mini mental state examination (MMSE)9

test scores and (b) the presence or absence ofdeficits in each of the neuropsychologicaldomains detailed in the NINCDS-ADRDAstatement (table 1)—namely, memory, orienta-tion, language, praxis, attention, perception,problem solving, and activities of daily livingand social function. In determining impair-ment, the rater adhered as closely as possible todefinitions of impairment given in the criteriastatement. Failure on one or more of thespecific cognitive tests recommended by thecriteria statement to tap a particular cognitivedomain was recorded as evidence of impair-

J Neurol Neurosurg Psychiatry 1999;66:184–188184

Department ofNeurologyA R VarmaJ S SnowdenP R TalbotD Neary

Department ofMedical Physics,Manchester RoyalInfirmary, Manchester,UKJ J Lloyd

Department ofPathological Sciences,Manchester MedicalSchool, Manchester,UKD M A Mann

Correspondence to:Dr AR Varma, CerebralFunction Unit, NeurologyDepartment, ManchesterRoyal Infirmary, OxfordRoad, Manchester M139WL, UK. Telephone 0044161 276 4138; fax 0044 161276 4681.

Received 28 April 1998 andin revised form9 July 1998Accepted 26 August 1998

on 9 June 2009 jnnp.bmj.comDownloaded from

ment in that domain. For example, a deficit inlanguage was scored as present if failureoccurred on any of the specified language testsincluding verbal fluency and naming tasks. Thescoring was conservative. If there was not wellestablished documented deficit then it wasscored absent. Whenever possible the reason forthe failure was also detailed.

In addition to the broad domains outlined bythe criteria, patients were scored with respectto the more narrowly defined variables ofprimary language, primary perception, andspatial domains (table 2). A deficit in primarylanguage was defined as impairment in one ormore of the linguistic domains of phonology,orthography, morphology, syntax, or lexicalsemantics. Evidence for impairment was basedon the presence of phonemic or semantic para-phasias in conversational speech, repetition ornaming tasks, of grammatical errors in sponta-neous speech, of impaired syntactic compre-hension, or paralexias in reading and spellingerrors in writing. Impoverished verbal fluencyor word retrieval, in the absence of phonologi-cal or semantic errors, was not construed asprimary language impairment. The primarylanguage variable was designed to tap impair-ments attributable to dysfunction in primarycortical language areas and exclude impair-ments on language tasks secondary to othercognitive deficits. Primary perceptual deficitwas defined as a problem in perceptual recog-nition of objects, line drawings, or faces. DiY-culty arising only on complex perceptual taskssuch as the Hooper visual test10 involving men-tal integration of visual information would beexcluded as diYculties on such complex taskscould arise for reasons outside the realm of pri-mary perception. Patients were designated tohave a spatial deficit if they had observablespatial disorientation as evidenced by diYcultyin negotiating the environment, orientingclothing when dressing, localisation of objects,or they failed in spatial tasks (tracking a maze,copying hand postures and line drawings, dotlocation, and dot counting tasks).

Twenty records were analysed by twoindependent raters (ARV and JSS) to assessinterrater reliability. There was good interrateragreement (ê range 0.6–0.8 for various do-mains). Missing values were generally low(median 2); this factor is therefore unlikely to

aVect the interrater reliability scores. The datafrom a total of 56 case records were availablefor final analysis. A record was judged to fulfilNINCDS-ADRDA criteria for probable AD if(a) the MMSE score was < 24 and (b) therewere demonstrable deficits in memory and atleast one other area of cognition as defined bythe criteria statement.

STATISTICS

For the NINCDS-ADRDA criteria and eachcognitive domain the sensitivity, specificity, andpositive likelihood ratio (LR) (with 95% confi-dence intervals) for probable AD were calcu-lated.

The sensitivity is the true positive fraction,the fraction of subjects with AD that have apositive test result (have a deficit in that cogni-tive domain). The specificity is the fraction ofsubjects that do not have AD and have a nega-tive test result (do not have that particular defi-cit). The LR11 for a positive test result (+LR) isdefined as the ratio of the fraction of subjectshaving a positive test result who have thedisease to the fraction having a positive testresult who do not have the disease. This can beexpressed as:

+LR= sensitivity1−specificity

Sensitivity and specificity are commonlyused indices of the usefulness of diagnostictests. However, they have the drawback thatthey cannot be easily applied to an individualpatient. LRs, by contrast, can be applied to anindividual patient and have an easily under-stood meaning—that is, modifying the pretestodds of disease. An LR of 1 indicates that noinformation has been provided by the test. Avalue> 1 indicates that the test increases theprobability of disease prevalence whereas avalue >1 decreases the probability of diseasepresence. LRs have the advantage that theyrelate the post-test odds of a disease to the pre-test odds in a simple way; post-test odds = +LR× pre-test odds. Odds of a disease is defined asthe ratio of the probability of having disease (P)to the probability of lack of disease; odds =P/(1-P). LRs were considered significant only ifthey diVered from 1 with 95% confidence.

Table 1 Patients showing deficits in individual cognitive domains specified by NINCDS-ADRDA criteria

Cognitive domain AD n (%) FTD n (%) Sensitivity Specificity LR 95% CI

Memory 28 (93.3) 20 (76.9) 0.93 0.23 1.21 0.96–1.53Orientation 26 (86.6) 12(46.1) 0.87 0.54 1.88 1.21–2.91*Language 28 (93.3) 24 (92.3) 0.93 0.08 1.01 0.87–1.17Praxis 28 (93.3) 11 (42.3) 0.93 0.58 2.21 1.39–3.49*Attention 24 (80) 19 (73.1) 0.8 0.27 1.09 0.82–1.47Perception 12 (40) 7 (26.9) 0.4 0.73 1.49 0.69–3.21Problem solving 12 (40) 18 (69.2) 0.4 0.31 0.58 0.35–0.96*

LR=positive likelihood ratio for AD; *LR significantly diVerent from 1.

Table 2 Patients showing deficits in cognitive domains not specified by NINCDS-ADRDA criteria

Cognitive domain AD n (%) FTD n (%) Sensitivity Specificity LR 95% CI

Primary language 22 (73) 12 (46.1) 0.73 0.54 1.59 1.00–2.54*Primary perception 9 (30) 1 (3.8) 0.3 0.96 7.8 1.06–57.53*Spatial 22 (73.3) 2 (7.69) 0.73 0.92 9.53 2.47–36.73*

LR=positive likelihood ratio for AD; *LR significantly diVerent from 1.

NINCDS-ADRDA in Alzheimer’s disease and frontotemporal dementia 185

on 9 June 2009 jnnp.bmj.comDownloaded from

ResultsOf the 56 patients with pathological confirma-tion of diagnosis, 30 had AD and 26 FTD. Theneuropathological characteristics of these pa-tients have been described previously.12–14 Themean age of onset and duration of illness atinitial referral were similar in the two groups(table 3). The mean age of the two groups inthis study reflect the referral pattern to aneurological unit. The sex distribution wasroughly equal in the AD group whereas malesoutnumbered females in the FTD group. Themean MMSE score was significantly higher inthe FTD group. In the AD group 29 of the 30patients had an MMSE score < 24 at the timeof initial diagnosis; one of these patients had nodemonstrable impairment of his memory.Hence, 28 of 30 patients were considered tofulfil NINCDS-ADRDA criteria for probableAD. In the FTD group 21 of the 26 patientshad an MMSE score<24; one of these patientshad no demonstrable memory impairment.Hence, 20 of 26 patients were considered tofulfil NINCDS-ADRDA criteria for probableAD. All patients were impaired in theiractivities of daily living and social function,which had led to their medical referral.

The overall sensitivity of the NINCDS-ADRDA criteria in diagnosing AD in thispopulation of patients with cortical dementia(AD or FTD) was 0.93; the specificity 0.23; theLR 1.21 (table 4). The sensitivity, specificity,and LRs for individual cognitive domains areshown in tables 1 and 2. The LR is significantlydiVerent from 1 if the value of 1 does not liewithin the 95% confidence interval. The fivedomains for which this was significantly greaterthan 1 were orientation, praxis, primarylanguage, and spatial and primary perception.For the domain of problem solving the LR wassignificantly less than 1. All the 20 patients withFTD who fulfilled NINCDS-ADRDA criteriafor probable AD would also fulfil currentLund-Manchester criteria for FTD.15

DiscussionThe NINCDS-ADRDA Work Group was setup to establish and describe clinical criteria forthe diagnosis of AD.7 The authors warned thatthese criteria were not yet fully operational andthat the criteria were to be regarded as tentativeand subject to change. Fourteen years laterthese criteria are still widely used in their origi-nal form. Patients who fulfil the NINCDS-ADRDA criteria are accepted as having a clini-

cal diagnosis of probable AD, a practicewitnessed in most research publications. Un-doubtedly these research practices permeatefrom published literature to clinical practice.

The validation of these criteria at necropsyhave been attempted in several studies, withvarying results.6 16–18 The accuracy rate (ratio ofthe number of pathologically confirmed cor-rect cases of AD to the number of clinicallydiagnosed cases of AD, expressed as thepercentage of clinical diagnosis of probable ADusing NINCDS-ADRDA ranges between 65%and 92%.6 However, these accuracy rates havebeen derived in all these studies for the diagno-sis of pathological AD, which includes ADalone and mixed cases (AD pathology plusvascular or Lewy bodies or Parkinson’s diseasepathology or other pathology). The rates fall ifthey are calculated using AD alone pathologyas the gold standard, and then vary between56% and 63%. Needless to say the accuracyrates are even lower for the clinical diagnosis ofpossible AD. NINCDS-ADRDA criteria arewidely applied in the diagnosis of AD. How-ever, the application of these criteria in the dif-ferentiation of two cortical dementias, AD andFTD, has never been studied before.

Both Mendez et al5 and Litvan et al19 havefound that Pick’s disease is often underdiag-nosed, the clinical misdiagnosis mainly attrib-utable to AD. It is clear, therefore, that FTD isunderdiagnosed and that AD is the commonestmisdiagnosis. The diagnostic diYculty is com-pounded because in both of these disorders thepatients are physically well and have few diVer-entiating neurological signs in the early stagesof the illness.20 The diagnosis mainly dependson a comprehensive neuropsychological evalu-ation. However, such assessment also requiresa careful qualitative analysis of cognition andbehaviour because patients can fail quantifiedpsychometric tests for various reasons.

This study shows that the NINCDS-ADRDA criteria have high sensitivity (0.93),but low specificity (0.23) in the diagnosis ofAD among a group of patients with corticaldementias (AD and FTD)—that is, althoughmost patients with AD fulfil clinical diagnosticcriteria for AD, so too do most patients withFTD. Furthermore, whether or not a patientfulfils the criteria does not significantly changethe odds of that patient having AD as opposedto FTD (the LR is not diVerent from 1 with95% confidence). These findings suggest thatthe ability of the criteria to discriminate ADand FTD is relatively poor. However, this studyhas also shown that the identification of deficitsin certain cognitive domains can contribute tothe clinical diVerentiation of AD and FTD;deficits in the realm of orientation in time andplace and praxis increase the odds of a patienthaving AD as compared to FTD, whereas defi-cits in problem solving decrease the odds of apatient having AD as opposed to FTD (table1). Deficits in the domains of attention,perception, language, and memory (as broadlydefined by the NINCDS-ADRDA statement)do not significantly change the odds of apatient having either condition (table 1).

Table 3 Demographic data

Group Sex (M:F)Age of onsetmean years (SD)

Duration of illnessmean years (SD)

MMSEmean (SD)

AD 14:16 55.7 (12.35) 3.29 (1.79) 10.69 (7.38)FTD 19:7* 56.5 (9.11) 2.98 (1.9) 19.78 (6.41)**

*p<0.05; **p<0.000. M=male; F=female; MMSE=mini mental state examination.

Table 4 Patients fulfilling NINCDS-ADRDA criteria for probable AD

NINCDS-ADRDAcriteria AD n FTD n Sensitivity Specificity LR 95% CI

Positive 28 20 0.93 0.23 1.21 0.96–1.53Negative 2 6

LR=positive likelihood ratio for AD.

186 Varma, Snowden, Lloyd, et al

on 9 June 2009 jnnp.bmj.comDownloaded from

Our patient groups (AD and FTD) weresimilar for their age of onset and duration ofillness at presentation (table 3). Thus anydiVerences found in their neuropsychologicaldeficits could not be accounted for by thesefactors. The patient population is young,reflecting the referral pattern to a neurologicalcentre. FTD is a disorder more commonly seenin the presenium.20 The clinical features seen inour patients are indistinguishable from thosepatients referred to a psychogeriatric centre.1

Patients with FTD had a significantly higherMMSE score than the patients with AD (table3). It is the more striking, given the lesser over-all severity of dementia as assessed by thisscore, that such a high percentage of patientswith FTD fulfil criteria for AD (table 4). Fail-ures in individual cognitive domains (tables 1and 2) across the two study groups (AD andFTD) are discussed below.

Memory impairment is virtually ubiquitousin dementing illnesses.21 The NINCDS-ADRDA criteria do not draw a distinctionbetween memory impairments due to a reten-tion disorder22 and those secondary to prob-lems in retrieval and organisation.23 The neuralsubstrates and hence the underlying causes forthese two types of memory dysfunction are dis-tinct. Patients with AD have a retention disor-der from pathology within their limbic struc-tures, whereas patients with FTD commonlyhave memory impairment due to retrieval andorganisational problems due to frontal lobedeficits. Both groups (AD and FTD) havememory deficit (as witnessed in this study),defined by impairment in memory tests, butfailure arises for widely diVerent reasons.These diVerences can be uncovered by appro-priate neuropsychological assessment andwould improve the diagnostic specificity of thecriteria.

The apparently poor discriminating power oflanguage impairment is probably a reflection ofthe relatively broad definition of linguistic dis-order outlined in the criteria statement. Thecriteria document suggests a wide range of ver-bal tasks potentially suitable for eliciting a lan-guage deficit, including verbal fluency and theBoston naming test. It is well established thatpatients with FTD show reduced verbalfluency and naming test performance as part ofa general adynamia and impoverishment inmotor responses. Indeed, verbal fluency iscommonly viewed as a test of frontal lobeexecutive function.24 Such broad criteria forlanguage disturbance will therefore inevitablyincorporate patients with FTD. When morenarrowly defined criteria for language impair-ment are adopted based on the presence ofimpairments in primary linguistic processesthen the diVerentiating value of this cognitivedomain increases substantially (tables 1 and 2).

Failure in attention tasks also has poordiscriminating value (table 1) between AD andFTD. This is not surprising because deficits inattention are a characteristic feature of FTD.The frontal lobes, primarily involved in FTDare known to have an important role inattention.25

The criteria statement suggests various tasksthat might be used to detect deficits in visualperception including Gollin incomplete pic-tures test and the Hooper test.10 Patients mayfail the Hooper complex perceptual tasks forreasons that are not primarily perceptual butarise for reasons of frontal lobe executivedysfunction such as attentional deficits or fail-ure in the ability to manipulate informationmentally. When more stringent criteria wereadopted to reflect a primary perceptual impair-ment, the LR improved fivefold (table 2),although few patients overall showed a deficitin this domain.

The identification of deficits in the realm ofpraxis significantly increases the odds of apatient having AD as opposed to FTD. TheNINCDS-ADRDA criteria indicate that fail-ure on drawing and block assembly tasksshould be considered a disorder of praxis.However, patients may fail these tests due todeficits in praxis or spatial function, or due topoor strategic and organisational skills (frontaldeficit). Although it is recognised that spatialabilities are commonly aVected in AD3 andstrikingly preserved in FTD,15 20 26 the criteriado not consider spatial function as a separatecognitive domain. Of the 11 (42%) patientswith FTD who failed in tests of praxis asdefined by the criteria, most failed due to defi-cits in strategy and organisation; only two (8%)patients failed specific spatial tasks (one faileddot counting tasks interpreted as secondary topoor sustained attention; the other did nottrack a maze due to impaired attention). Bycontrast 22 (73.3%) patients with AD had spa-tial impairment (LR 9.53, table 2), indicatinghigh discriminating value for this domain. Theidentification of a disorder of spatial functionas strictly defined increases considerably theodds of a patient having AD as opposed toFTD (table 2).

Twenty (77%) patients with pathologicalconfirmation of FTD in this study fulfilledNINCDS-ADRDA criteria for AD. Eighteen(70%) of these patients fulfilled criteria on thebasis of failure in the domains of memory,attention, and problem solving only. In view ofthe fact that the criteria require deficits in onlytwo cognitive domains (one of which ismemory) and the finding that failure inproblem solving tasks significantly decreases theodds of a patient having AD as opposed toFTD, it would seem reasonable to consider theexclusion of problem solving tasks from thecriteria.

It has been suggested that formal psycho-metric testing is rarely necessary for thediagnosis of AD even though NINCDS-ADRDA criteria called for confirmation of thedementia by neuropsychological testing.3 Thefindings from the present study suggest, on thecontrary, the need for more careful documen-tation of cognitive impairments. Such an exer-cise should be aimed at identifying not merelytest failure, but in discerning the nature of thecognitive deficits underlying such failure, asthis has implications for its neural substrateand for clinical diagnosis. A revision of existingNINCDS-ADRDA criteria involving more

NINCDS-ADRDA in Alzheimer’s disease and frontotemporal dementia 187

on 9 June 2009 jnnp.bmj.comDownloaded from

narrowly defined definitions of cognitive defi-cits ought to improve diagnostic specificitywithout loss of overall sensitivity. It is suggestedthat memory be considered impaired only if thepatient is also disoriented in time and place.The amalgamation of memory and orientationinto one cognitive domain would improve spe-cificity (table 1). A further characterisation oflanguage and perception as primary languageand primary perception will improve specificitytoo. The addition of spatial function andremoval of problem solving and attentionshould also strengthen the criteria. Because amajority of patients with FTD fulfil theNINCDS-ADRDA criteria it is further sug-gested that those patients who fulfil thediagnosis of probable AD, but who, in addition,fulfil the published criteria for FTD,15 shouldbe viewed with diagnostic caution and notconfidently considered to have probable AD.

FTD is normally the only non-AD form ofdementia that can lead to diagnostic confusion.It has been shown that patients with Lewy bodydementia (LBD) may also fulfil NINCDS-ADRDA criteria for AD.27 The advent of neu-roimaging including SPECT,28 and betterhistopathological and immunocytochemicaltechniques have led to the awareness of diVer-ent neurodegenerative disorders causing de-mentia and which have characteristic neuro-psychological profiles. When the NINCDS-ADRDA criteria were first formulated theirmain aim was to diVerentiate between AD andvascular dementia. Any future revisions andrefinements of the present criteria will need totake into account the diversity of neurodegen-erative disorders leading to dementia whichhave neither the neuropsychological profile northe characteristic pathology of AD.

We are grateful to the carers of our patients for donating thebrains of their relatives for research.

1 Gustafson L. Frontal lobe degeneration of non-Alzheimertype. II. Clinical picture and diVerential diagnosis. ArchGerontol Geriatr 1987;6:209–23.

2 Neary D, Snowden JS, Northen B, et al. Dementia of frontallobe type. J Neurol Neurosurg Psychiatry 1988;51:353–61.

3 Berg L, Morris JC. Diagnosis. In: Terry RD, Katzman R,Bick KL, eds. Alzheimer disease. New York: Raven Press,1994:9–25.

4 Trimble MR. The dementias. In: Trimble MR. Biologicalpsychiatry. Chichester: John Wiley, 1996:291–324.

5 Mendez MF, Selwood A, Mastri AR, et al. Pick’s diseaseversus Alzheimer’s disease: a comparison of clinical charac-teristics. Neurology 1993;43:289–92.

6 Klatka LA, SchiVer RB, Powers JM, et al. Incorrect diagno-sis of Alzheimer’s disease, a clinicopathological study. ArchNeurol 1996;53:35–42.

7 McKhann G, Drachman D, Folstein M, et al. Clinical diag-nosis of Alzheimer’s disease: report of the NINCDS-ADRDA Work Group under the Auspices of Departmentof Health and Human Services Task Force on Alzheimer’sdisease. Neurology 1984;34:939–44.

8 Neary D, Snowden JS, Bowen DM, et al. Neuropsychologi-cal syndromes in presenile dementia due to cerebralatrophy. J Neurol Neurosurg Psychiatry 1986;49:163–74.

9 Folstein MF, Folstein SE. Mini-mental state: a practicalmethod for grading the cognitive state of patients for theclinician. J Psychiatr Res 1975;12:189–98.

10 Hooper HE. The Hooper visual organisation test. Los Angeles:Western Psychological Services, 1958.

11 Black WC, Armstrong P. Communicating the significance ofradiologic test results: the likelihood ratio. American Journalof Radiology 1986;147:1313–8.

12 Mann DMA. The topographical distribution of brainpathology in Alzheimer’s disease. Acta Neuropathol (Berl)1991;83:81–6.

13 Mann DMA, South PW, Snowden JS, et al. Dementia offrontal lobe type; neuropathology and immunohistochem-istry. J Neurol Neurosurg Psychiatry 1993;56:605–14.

14 Mann DMA, South PW. The topographic distribution ofbrain atrophy in frontal lobe dementia. Acta Neuropatho-logica 1993;83:334–40.

15 Brun A, Englund B, Gustafson L, et al. Clinical andneuropathological criteria for frontotemporal dementia.The Lund and Manchester Groups. J Neurol NeurosurgPsychiatry 1994;57:416–8.

16 Galasko D, Hansen LA, Katzman R, et al. Clinical-neuropathological correlations in Alzheimer’s disease andrelated dementias. Arch Neurol 1994;51:888–95.

17 Kukull WA, Larson EB, Reifler BV, et al. The validity ofthree clinical diagnostic criteria for Alzheimer’s disease.Neurology 1990;40:1364–9.

18 Boller F, Lopez OL, Moossy J. Diagnosis of dementia: clin-icopathological correlations. Neurology 1989;39:76–9.

19 Litvan I, Agid Y, Sastrj N, et al. What are the obstacles for anaccurate clinical diagnosis of Pick’s disease? A clinico-pathological study. Neurology 1997;49:62–9.

20 Snowden JS, Neary D, Mann DMA. Fronto-temporaldementia. In: Fronto-temporal lobar degeneration: fronto-temporal dementia, progressive aphasia, semantic dementia.New York: Churchill Livingstone, 1996:9–41.

21 Snowden JS. Contribution to the diVerential diagnosis ofdementias.1: Neuropsychology. Reviews in Clinical Geron-tology 1994;4:227–34.

22 Scoville WB, Milner B. Loss of recent memory after bilateralhippocampal lesions. J Neurol Neurosurg Psychiatry 1957;20:11–21.

23 Moss MB, Albert MS, Butters N, et al. DiVerential patternsof memory loss among patients with Alzheimer’s disease,Huntington’s disease and alcoholic KorsakoV’s syndrome.Arch Neurol 1986;43:239–46.

24 Milner B. Some eVects of frontal lobectomy in man. In: JMWarren, K Abert, eds. The frontal granular cortex and behav-iour. New York: McGraw Hill, 1964:313–31.

25 Stuss DT, Benson FD. Attention. In: Stuss DT, Benson FD,eds. The frontal lobes. New York: Raven Press, 1986:97–106.

26 Barber R, Snowden JS, Craufurd D. Frontotemporaldementia and Alzheimer’s disease: retrospective diVeren-tiation using information from informants. J Neurol Neuro-surg Psychiatry 1995;59:61–70.

27 McKeith IG, Fairbairn AF, Perry RF, et al. The clinicaldiagnosis and misdiagnosis of senile dementia of Lewybody type (SDLT). Br J Psychiatry 1994;165:324–32.

28 Talbot PR, Snowden JS, Lloyd JJ, et al. The contribution ofsingle photon emission tomography to clinical diVerentia-tion of degenerative cortical brain disorders. J Neurol 1995;242:579–86.

188 Varma, Snowden, Lloyd, et al

on 9 June 2009 jnnp.bmj.comDownloaded from