5
 T he many cognitive screening instruments in existence may be broadly classified as scales that test either general or specific cog- nitive functions. 1 The former, including the Mini-Mental State Examination (MMSE), 2 th e  Add enbr ooke ’s Cognitiv e Exam - ination (ACE) 3 and its revision (ACE-R), 4 and the Montreal Cognitive Assessment (MoCA), 5 have been used most widely in the assessment of patients with possi- ble dementia. However, scales that test a specific cognitive function, such as memory, 6 may also have a place in clinical practice, since these tests are usually very short, some requiring only one or two questions. 7-9  Visual percep tual deficit s ma y be a feature of various dementia syndromes, including Alzheimer’s disease, especially in its ‘visual vari- ant’, also known as posterior corti- cal atrophy, 10 in Parkinson’s disease dementia and dementia  with Lewy bodies, and in prion dis- orders. Frontotemporal lobar degenerations, by contrast, classi- cally spare visual perceptual func- tion, although Luria noted that  patients with frontal lobe lesions  were cha rac ter ise d by per sis ten t inert attention to only one detail of a pattern, reflecting an inabil- ity to move from one component to another as easily as they should. 11 The MMSE is recognised to lack a specific test of visual percep- tual function, the intersecting pentagons being a test of visuo - motor or visuoconstructional function. This is also true of the  AC E (N ec ker cu be , clock dr aw - ing) and the ‘visuospatial/execu- tive’ tests in the MoCA. The  ACE-R, however , has specific visual perceptual tasks (dot counting, fragmented letters). Overlapping figures used in the assessment of visual perceptual deficits comprise two or more line drawings that partly overlap. 12 The German neuropsychiatrist Walter Poppelreuter (1886-1939) 13 pro- duced one example of overlapping figures (see Figure 1), 14 as part of his interest in the study of visual neglect, which he believed was a form of inattention. 15 Th e Poppelreuter figure, sometimes known as the Ghent or Poppelreuter-Ghent figure, may be characterised (in Gestalt terms) as a figure/ground discrimination task, which is acknowledged to be problematic for patients with apperceptive (but not associative)  vi sual agn osi a, es pe ci al ly do rs al simultanagnosia, usually associated  wit h r igh t p ost eri or hemisp her e lesions. 16,17 Despite its antiquity, the clini- cal utility of the Poppelreuter fig- ure, as opposed to its value in neuropsychologi cal studies, has seldom been examined. In a group of 12 Alzheimer’s disease patients, Della Sala et al  . found that performance was impaired using their overlapping figure paradigm, principally at the figure discrimination stage, and con- cluded that the Poppelreuter fig- ure was a sensitive tool to detect  visu ospa tial defi cits. By cont rast , the only errors made by healthy individuals were omissions. 18 The aim of this study was to assess the utility of the Poppelreuter figure for the diag- nosis of dementia in day-to-day clinical practice in the setting of a dedicated cognitive function clinic, using both retrospective and prospective assessments. Patients and methods The retrospective part of the study  was p erformed by e xamining the notes of 50 patients administered the Poppelreuter figure from 2001- Progress in Neurology and Psychiatry 17 www.progressnp.com Poppelreuter figure   Original research Figure 1. The Poppelreuter figure, consisting of four overlapping images, is a simple test of visual per ceptual function The Poppelreuter figure visual perceptual function test for dementia diagnosis Rebecca Sells, Andrew J Larner MRCP The Poppelreuter figure is a simple test of visual perceptual function consisting of four overlapping images. This study investigates its utility for the diagnosis of dementia in a cognitive function clinic.

The Poppelreuter Figure Visual Perceptual

  • Upload
    icaro

  • View
    27

  • Download
    0

Embed Size (px)

DESCRIPTION

Poppelreuter

Citation preview

  • The many cognitive screeninginstruments in existence maybe broadly classified as scales thattest either general or specific cog-nitive functions.1 The former,including the Mini-Mental StateExamination (MMSE),2 theAddenbrookes Cognitive Exam -ination (ACE)3 and its revision(ACE-R),4 and the MontrealCognitive Assessment (MoCA),5

    have been used most widely in theassessment of patients with possi-ble dementia. However, scales thattest a specific cognitive function,such as memory,6 may also have aplace in clinical practice, sincethese tests are usually very short,some requiring only one or twoquestions.7-9

    Visual perceptual deficits maybe a feature of various dementiasyndromes, including Alzheimersdisease, especially in its visual vari-ant, also known as posterior corti-cal atrophy,10 in Parkinsonsdisease dementia and dementiawith Lewy bodies, and in prion dis-orders. Frontotemporal lobardegenerations, by contrast, classi-cally spare visual perceptual func-tion, although Luria noted thatpatients with frontal lobe lesionswere characterised by persistentinert attention to only one detailof a pattern, reflecting an inabil-ity to move from one componentto another as easily as theyshould.11

    The MMSE is recognised tolack a specific test of visual percep-

    tual function, the intersectingpentagons being a test of visuo -motor or visuoconstructionalfunction. This is also true of theACE (Necker cube, clock draw-ing) and the visuospatial/execu-tive tests in the MoCA. TheACE-R, however, has specific visualperceptual tasks (dot counting,fragmented letters).

    Overlapping figures used in theassessment of visual perceptualdeficits comprise two or more linedrawings that partly overlap.12 TheGerman neuropsychiatrist WalterPoppelreuter (1886-1939)13 pro-duced one example of overlappingfigures (see Figure 1),14 as part ofhis interest in the study of visualneglect, which he believed was aform of inattention.15 ThePoppelreuter figure, sometimesknown as the Ghent orPoppelreuter-Ghent figure, may becharacterised (in Gestalt terms) asa figure/ground discriminationtask, which is acknowledged to beproblematic for patients withapperceptive (but not associative)visual agnosia, especially dorsalsimultanagnosia, usually associatedwith right posterior hemispherelesions.16,17

    Despite its antiquity, the clini-cal utility of the Poppelreuter fig-ure, as opposed to its value inneuropsychological studies, hasseldom been examined. In agroup of 12 Alzheimers diseasepatients, Della Sala et al. foundthat performance was impaired

    using their overlapping figureparadigm, principally at the figurediscrimination stage, and con-cluded that the Poppelreuter fig-ure was a sensitive tool to detectvisuospatial deficits. By contrast,the only errors made by healthyindividuals were omissions.18

    The aim of this study was toassess the utility of thePoppelreuter figure for the diag-nosis of dementia in day-to-dayclinical practice in the setting of adedicated cognitive functionclinic, using both retrospective andprospective assessments.

    Patients and methodsThe retrospective part of the studywas performed by examining thenotes of 50 patients administeredthe Poppelreuter figure from 2001-

    Progress in Neurology and Psychiatry 17www.progressnp.com

    Poppelreuter figure z Original research

    Figure 1. The Poppelreuter figure, consisting of four overlappingimages, is a simple test of visual perceptual function

    The Poppelreuter figure visual perceptualfunction test for dementia diagnosisRebecca Sells, Andrew J Larner MRCP

    The Poppelreuter figure is a simple test of visual perceptual function consisting of fouroverlapping images. This study investigates its utility for the diagnosis of dementia in acognitive function clinic.

    Original_Poppelreuter_Layout 1 05/04/2011 15:13 Page 1

  • 2002. In the prospective part of thestudy, consecutive new patientreferrals were recruited over a six-month period (March-September2010). In both parts of the study,standard clinical diagnostic crite-ria were used for the diagnosis ofdementia (DSM-IV) and dementiasubtypes, as previously described.19

    As the prospective evaluation wasa pragmatic assessment, patientswere selected as they presented tothe cognitive function clinic, notaccording to diagnosis. Diagnosiswas by the judgement of an experi-enced clinician based on diagnos-tic criteria.

    In both the retrospective andprospective assessments, patientswere shown the Poppelreuter figure(black and white line drawing,image size 9cm x 9cm; see Figure 1),always presented initially in thesame orientation (namely with thejug vertical), independent of, buton the same day as, clinical and neu-ropsychological assessment. ThePoppelreuter figure test was scored(range 0-4) according to thepatients ability to identify, not toname, the individual elements (jug,cleaver, iron, hammer). Points werededucted for omissions, but not forfalse alarms, ie naming items notpresent. Poppelreuter scores werenot used in the diagnostic judgmentof dementia/no dementia in orderto minimise review bias.20

    Other tests administered werethe MMSE,2 ACE3 (2001-2002study only), and the two tests ofvisual perceptual abilities from theACE-R4 (2010 study only), namelycounting of dot arrays and identi-fication of fragmented letters.Correlations were calculatedbetween the scores for thePoppelreuter figure and totalMMSE score (range 0-30), MMSEintersecting pentagons (0-1), ACEtotal score (0-100), ACE-R dotcounting (0-4), and ACE-R frag-mented letters (0-4). Standard

    summary measures of diagnosticutility (for dementia vs no demen-tia) were generated for Poppel -reuter figure scores; namely,sensitivity, specificity, Youdenindex, positive and negative pre-dictive values (PPV, NPV), diagnos-tic odds ratio, positive and negativelikelihood ratios (LR+ and LR-),positive and negative utility index(UI+, UI-),21 and receiver operat-ing characteristic (ROC) curve.The Standards for the Reportingof Diagnostic Accuracy (STARD)guidelines on reporting diagnostictest accuracy were observed.22

    Results Retrospective cohortOverall, 50 patients were assessedwith the Poppelreuter figure, ofwhom 28 (56 per cent) haddementia by DSM-IV criteria (seeTable 1, left-hand column, for

    demographics and diagnoses). Norecord of patient acceptability of the Poppelreuter figure wasrecorded in the notes.

    Poppelreuter scores rangedfrom 0-4. For the group withdementia, the mode, median andmean Poppelreuter scores were 4,4, and 3.50 0.72, respectively. Allthe patients without dementiascored 4. The mean Poppelreuterscores differed significantly betweenthe groups with and without demen-tia (t = 3.09, df = 42, p < 0.01).

    There was a weak negative correlation between age andPoppelreuter score (r = -0.23).Correlations between Poppel -reuter score and the other testsshowed a high correlation only forintersecting pentagons score (r =0.80, p < 0.001), the other correla-tions being low (MMSE: r = 0.26, p < 0.05; ACE: r = 0.23, p < 0.05).

    Original research z Poppelreuter figure

    Progress in Neurology and Psychiatry18 www.progressnp.com

    Demographics: Retrospective Prospective cohortcohort (2001-2002) (March-September 2010)

    Number 50 101

    M:F ratio 35:15 (70% male) 53:48 (52 % male)

    Age range 41-82 years, 23-89 years, median 60 years median 61 years

    Dementia:no 28:22 (dementia 28:73 (dementia dementia ratio prevalence = 56%) prevalence = 28%)

    Dementia diagnoses

    Alzheimers disease 20 (1 posterior 15 (2 posterior cortical cortical atrophy) atrophy)

    Frontotemporal 7 8lobar degeneration

    Dementia with 1 3Lewy bodies/Parkinsons disease dementia

    Others 0 2

    Table 1. Demographics and diagnoses of the patients in the retrospective and prospectivecohorts

    Original_Poppelreuter_Layout 1 05/04/2011 15:13 Page 2

  • Optimal test sensitivity for thedifferential diagnosis of demen-tia/not dementia in this cohortwas found to be 0.38 at thePoppelreuter cutoff of 3/4, andoptimal test specificity was 1.00 atthe same cutoff, with test accuracyof 0.66. Traditional parameters oftest diagnostic utility at the 3/4cutoff are shown in Table 2 (left-hand column).

    Prospective cohortOver the study period, 101 patientswere assessed of whom 28 (28 percent) had dementia as assessed byDSM-IV criteria (see Table 1, righthand column, for demographicsand diagnoses). The Poppelreuterfigure proved acceptable topatients and easy to use, beingcompleted in less than one minuteby all patients.

    Poppelreuter scores rangedfrom 0-4. For the group withdementia, the mode, median andmean Poppelreuter scores were 4,4, and 3.32 1.09, respectively; forthe group without dementia themode, median, and mean scoreswere 4, 4, and 3.85 0.36. Themean Poppelreuter scores differedsignificantly between the groups

    with and without dementia (t = 3.67, df = 99, p < 0.001).

    There was a very weak negativecorrelation between age andPoppelreuter score (r = -0.13).Correlations between Poppel -reuter score and the other specific(dot counting, fragmented letters,intersecting pentagons) and gen-eral (MMSE) tests were moderateor high for other visual perceptualtasks (dot counting: r = 0.56, p < 0.001; fragmented letters: r = 0.71, p < 0.001) and poor forthe visuomotor task (intersectingpentagons: r = 0.23, p < 0.05) andfor MMSE (r = 0.45, p < 0.001).

    Optimal test sensitivity for thedifferential diagnosis of demen-tia/not dementia in this cohort wasfound to be 0.39 at the Poppel -reuter cutoff of 3/4, and optimaltest specificity was 1.00 at the cut-off of 2/4, with similar test accu-racy at both cutoffs (0.72, 0.77respectively). Traditional parame-ters of test diagnostic utility at the 3/4, cutoff are shown in Table 2(right-hand column). Of particularnote, the utility indices indicatedthat the Poppelreuter figure wasmore useful for ruling out a diag-nosis of dementia (good negative

    utility index) than for ruling it in(very poor positive utility index).

    DiscussionThe Poppelreuter figure provedeasy to use in clinical practice andacceptable to patients. This mayrelate in part to the relatively lowpercentage of dementia patientsseen in the prospective cohort (28per cent), half that in the retro-spective cohort (albeit not consec-utive patients), a trend recentlyobserved in this clinic and possi-bly related to the recommenda-tions of the National DementiaStrategy prompting referral ofmore worried well memory com-plainers.23

    Although considered a multi-componential task by someauthors, the visuospatial compo-nents of the Poppelreuter figureare clearly the most important.18

    It was therefore not surprising thatthe Poppelreuter results were cor-related with other tests of visualperceptual function, particularlythe fragmented letters test, whichrequires figure/ground discrimi-nation, in comparison to generalscreening tests such as MMSE andACE, hence showing concurrentvalidity. The reason(s) for the dif-ference in Poppelreuter correla-tion with the visuoconstructionalintersecting pentagons taskbetween the two cohorts is uncer-tain, but may possibly be related tocase mix.

    The Poppelreuter figureproved very specific but not verysensitive for a diagnosis of demen-tia in both cohorts, and the overalldiagnostic accuracy as measuredby the ROC curve was suboptimal(> 0.75 is usually considered desir-able for a diagnostic test), particu-larly when compared to somegeneral screening tests such as theACE-R,24 but on a par with scalessuch as the PHQ-9.25 However,like the PHQ-9, the Poppelreuter

    Original research z Poppelreuter figure

    Progress in Neurology and Psychiatry20 www.progressnp.com

    Retrospective Prospective cohort cohort (March-(2001-2002) September 2010)

    Test accuracy 0.66 (0.52-0.80) 0.72 (0.64-0.81)Sensitivity 0.38 (0.18-0.57) 0.39 (0.21-0.57)Specificity 1.00 0.85 (0.77-0.93)Youden index (sens + spec - 1) 0.38 0.24Positive predictive value (PPV) 1.00 0.50 (0.29-0.71)Negative predictive value (NPV) 0.57 (0.41-0.74) 0.78 (0.69-0.88)Diagnostic Odds Ratio 3.65 (1.79-7.44)Positive likelihood ratio (LR+) 2.61 (1.28-5.32)Negative likelihood ratio (LR-) 0.63 (0.37-1.05) 0.71 (0.35-1.46)Utility index (UI+) 0.38 (poor) 0.20 (very poor)Utility index (UI-) 0.57 (satisfactory) 0.67 (good)Area under ROC curve 0.69 (0.59-0.79) 0.63 (0.53-0.74)

    Table 2. Diagnostic parameters for Poppelreuter test at cutoff 3/4 (with 95% confidenceintervals)

    Original_Poppelreuter_Layout 1 05/04/2011 15:13 Page 3

  • Progress in Neurology and Psychiatry 21www.progressnp.com

    Poppelreuter figure z Original research

    figure has pragmatic value: thegood negative utility index indi-cates that it is useful for ruling outa diagnosis of dementia, ratherthan for ruling it in (poor positiveutility index). A possible excep-tion to this rule may be cases ofposterior cortical atrophy (visualvariant Alzheimers disease): allthe cases in these studies per-formed below the cutoff 3/4.

    Although numbers of caseswere too small to permit examina-tion of test performance with stageof dementia, the low sensitivityresult suggests the Poppelreuterfigure would not be a useful screenfor early identification of demen-tia, but poor performance is a reli-able indicator that dementia hasset in.

    Nonetheless, because of thegood negative utility index, thePoppelreuter figure is useful forruling out a diagnosis of demen-tia. The Poppelreuter figuremight therefore be useful as avisual perceptual task in a generaldementia screening test or as onecomponent of a broader assess-ment battery. It might also proveto be a useful and quick stand-alone screen for dementia, per-haps readily applicable in primarycare where time available for test-ing is brief.

    Rebecca Sells is a final year medicalstudent and Dr Larner is a

    Consultant Neurologist, CognitiveFunction Clinic, Walton Centre forNeurology and Neurosurgery,Liverpool

    Declaration of interestsNone declared.

    References1. Tate RL. A Compendium of Tests, Scales, andQuestionnaires. The practitioners guide tomeasuring outcomes after acquired brainimpairment. Hove: Psychology Press, 2010.2. Folstein MF, Folstein SE, McHugh PR. Mini-Mental State. A practical method for gradingthe cognitive state of patients for the clinician.J Psychiatr Res 1975;12:189-98.3. Mathuranath PS, Nestor PJ, Berrios GE, etal. A brief cognitive test battery to differenti-ate Alzheimers disease and frontotemporaldementia. Neurology 2000;55:1613-20.4. Mioshi E, Dawson K, Mitchell J, et al. TheAddenbrookes Cognitive Examination Revised(ACE-R): a brief cognitive test battery fordementia screening. Int J Geriatr Psychiatry2006;21:1078-85.5. Nasreddine ZS, Phillips NA, Bdirian V, et al.The Montreal Cognitive Assessment, MoCA: abrief screening tool for mild cognitive impair-ment. J Am Geriatr Soc 2005;53:695-9.6. Buschke H, Kuslansky G, Katz M, et al.Screening for dementia with the MemoryImpairment Screen. Neurology 1999;52:231-8.7. St John P, Montgomery P. Is subjective mem-ory loss correlated with MMSE scores ordementia? J Geriatr Psychiatry Neurol2003;16:80-3.8. Chong MS, Chin JJ, Saw SM, et al. Screeningfor dementia in the older Chinese with a sin-gle question test on progressive forgetfulness.Int J Geriatr Psychiatry 2006;21:442-8.9. Larner AJ. Attended alone sign: validityand reliability for the exclusion of dementia.Age Ageing 2009;38:476-8.10. Mendez MF, Ghajarania M, Perryman KM.Posterior cortical atrophy: clinical characteris-

    tics and differences compared to Alzheimersdisease. Dement Geriatr Cogn Disord 2002;14:33-40.11. Luria AR. Higher Cortical Function in Man(2nd edn). New York: Basic Books, 1980:329.12. Loring DW, ed. INS Dictionary ofNeuropsychology. New York: OxfordUniversity Press, 1999:119.13. Fink GR. Walter Poppelreuter (1886-1939).Remarks on the cover illustration [in German].Nervenarzt 2003;74:540-1.14. Poppelreuter W. Die psychischenSchdigungen durch Kopfschuss im Kriege1914/17: mit besonderer Bercksichtigung derpathopsychologischen, pdagogischen,gewerblichen und sozialen Beziehungen (2 vol-umes. Band 1: Die Strungen der niederen undhheren Sehleistungen durch Verletzungendes Okzipitalhirns; Band 2: Die Herabsetzungder krperlichen Leistungsfhigkeit und desArbeitswillens durch Hirnverletzung imVergleich zu Normalen und Psychogenen).Leipzig: Voss, 1917-1918.15. Poppelreuter W. Zur Psychologie undPathologie der optischen Wahrnehmung.Zeitschrift fr die Gesamte Neurologie undPsychiatrie 1923;83:26-152.16. De Renzi E, Scotti G, Spinnler H. Perceptualand associative disorders of visual recognition.Relationship to the side of the cerebral lesion.Neurology 1969;19:634-42.17. Shallice T. From Neuropsychology toMental Structure. Cambridge: CambridgeUniversity Press, 1988:189-90.18. Della Sala S, Laiacona M, Trivelli C,Spinnler H. Poppelreuter-Ghents overlappingfigures test: its sensitivity to age and its clin-ical use. Arch Clin Neuropsychol 1995;10:511-34.19. Hancock P, Larner AJ. Test Your Memory(TYM) test: diagnostic utility in a memory clinicpopulation. Int J Geriatr Psychiatry 2011;26:inpress.20. Gifford DR, Cummings JL. Evaluatingdementia screening tests. Methodologic stan-dards to rate their performance. Neurology1999;52:224-7.21. Mitchell AJ, McGlinchey JB, Young D, et al.Accuracy of specific symptoms in the diagno-sis of major depressive disorder in psychiatricout-patients: data from the MIDAS project.Psychol Med 2009;39:1107-16.22. Bossuyt PM, Reitsma JB, Bruns DE, et al.The STARD statement for reporting studies ofdiagnostic accuracy: explanation and elabora-tion. Clin Chem 2003;49:7-18.23. Larner AJ. Impact of the NationalDementia Strategy in a neurology-led mem-ory clinic. Clin Med 2010;10:526.24. Larner AJ. Addenbrookes CognitiveExamination-Revised (ACE-R) in day-to-dayclinical practice. Age Ageing 2007;36:685-6.25. Hancock P, Larner AJ. Clinical utility ofPatient Health Questionnaire-9 (PHQ-9) inmemory clinics. Int J Psychiatry Clin Pract2009;13:188-91.

    Key points

    Visual perceptual function may be impaired in cognitive disorders,including Alzheimers disease

    Many of the standard cognitive screening instruments include little orno testing of visual perceptual function

    The Poppelreuter figure of four overlapping images is a simple test ofvisual perceptual function, easily applicable in the clinic and acceptableto patients

    In both retrospective and prospective studies in cognitive clinic patients,the Poppelreuter figure had a high specificity for the diagnosis ofdementia, scores above a cutoff of 3/4 ruling out a diagnosis ofdementia

    Original_Poppelreuter_Layout 1 05/04/2011 15:13 Page 4