Embed Size (px)
Citation preview
1446
Anosognosia in Patients WithCerebrovascular LesionsA Study of Causative Factors
Sergio E. Starkstein, MD, PhD; J. Paul Fedoroff, MD; Thomas R. Price, MD;Ramon Leiguarda, MD; and Robert G. Robinson, MD
Background and Purpose: Psychological and biological hypotheses have been proposed to explainanosognosia. We correlated the presence of anosognosia with the presence and severity of psychiatricdisturbances, neglect, intellectual impairments, and computed tomographic evidence of lesion size,location, and measurements of brain atrophy.
Methods: A series of 80 patients with acute stroke were assessed using a battery of psychiatric andneuropsychological tests and computed tomography.
Results: There were five main findings. First, 27 (28%) of the 96 patients originally screened showedanosognosia. Second, patients with anosognosia had significantly higher frequencies of hemispatialneglect and related phenomena, as well as deficits in recognizing facial emotions and in receptive prosody.Third, depression was equally frequent among patients with and without anosognosia. Fourth, patientswith anosognosia had a significantly higher frequency of right hemisphere lesions, primarily involving thetemporoparietal junction, thalamus, and basal ganglia. Fifth, patients with anosognosia showed signifi-cantly more subcortical brain atrophy, primarily involving the frontal white matter and diencephalicareas.
Conclusions: The present study demonstrates that anosognosia does not "protect" stroke patients fromdepressive feelings; rather, it represents arousal-attentional disorders after lesions in specific areas of theright hemisphere in nonaphasic patients with preexisting subcortical atrophy. (Stroke 1992;23:1446-1453)
KEYWORDS • anosognosia • depression
A nosognosia is a term used by Babinski1 to refer to/ \ patients who did not recognize their hemiple-
X A . gia. The term was later used to describe thedenial of other deficits, such as blindness (Anton'ssyndrome), hemianopsia, etc.2 Two main hypotheseshave been advanced to explain the presence of anosog-nosia. The psychological hypothesis suggests thatanosognosia is the consequence of a defensive mecha-nism by which patients protect themselves from realiz-ing the full extent of their deficits.3 The biologicalhypothesis postulates several factors that may underlie
From the Department of Psychiatry (S.E.S., J.P.F., R.G.R.),Johns Hopkins University School of Medicine, Baltimore, Md.; theDepartment of Psychiatry (S.E.S., R.G.R.), University of IowaCollege of Medicine, Iowa City, Iowa; the Department of Behav-ioral Neurology (S.E.S., R.L.), Institute of Neurological ResearchRaul Carrea, Buenos Aires, Argentina; the Department of Neu-rology (T.R.P.), University of Maryland School of Medicine,Baltimore, Md.; and the Division of Forensic Psychiatry (J.P.F.),The Clarke Institute of Psychiatry, Toronto, Canada.
Supported in part by National Institutes of Health grantsNS-15080 and MH-40355, research scientist award MH-00163 (toR.G.R.), a grant from the Instituto Di Telia (to S.E.S.), and aNational Alliance for Research in Schizophrenia and DepressionYoung Investigator Award (to S.E.S.).
Address for reprints: Sergio E. Starkstein, MD, PhD, Depart-ment of Behavioral Neurology, Institute of Neurological ResearchRaul Carrea, Ayacucho 2170, 1112 Buenos Aires, Argentina.
Received December 27, 1991; final revision received May 18,1992; accepted May 19, 1992.
the presence of anosognosia, such as lesion location(anosognosia may be more frequent after lesions of theright inferior parietal cortex), neglect (anosognosia maybe a manifestation of hemi-inattention), or intellectualdeficits (anosognosia may be significantly more frequentamong patients with cognitive deficits).4
If the psychological hypothesis is correct, a doubledissociation between anosognosia and depressionshould be expected (i.e., anosognosia should "protect"patients from depressive feelings produced by the newlyacquired neurological deficit). On the other hand, if thebiological hypothesis is correct, there might be anassociation between anosognosia and lesion location,neglect, or intellectual impairments.
To examine the mechanism of anosognosia, we stud-ied a series of patients with acute cerebrovascularlesions for the presence of anosognosia and correlatedits presence and severity with psychiatric disturbances(depression, anxiety, and apathy), neglect, intellectualimpairments, and computed tomographic (CT) evi-dence of lesion size, location, and measurements ofbrain atrophy.
The focus of our study was to examine whetheranosognosia 1) protects stroke patients from depressivefeelings, 2) is significantly associated with neglect andrelated phenomena, 3) is significantly related to cogni-tive deficits, 4) is significantly associated with lesionvariables such as side, size, and location, and 5) is
by guest on June 29, 2018http://stroke.ahajournals.org/
Dow
nloaded from
Starkstein et al Stroke and Anosognosia 1447
significantly related to either cortical or subcorticalbrain atrophy.
Subjects and MethodsPatients
Patients included in this study were selected from 96consecutive admissions to the University of MarylandHospital, Baltimore, for treatment of acute thromboem-bolic infarction or intracerebral hemorrhage. Patientswith a history of cerebrovascular lesions (« = 10) ormoderate to severe comprehension deficits on the shortform of the Token Test5 («=6) were excluded from thestudy. Patients with nonfluent aphasia, however, whocould reliably answer questions with affirmative ornegative answers were included.
Neurological ExaminationThe neurological examination and diagnosis were
made using the procedures and criteria established bythe Stroke Data Bank of the National Institute ofNeurological Disorders and Stroke, Bethesda, Md.6 Allneurological evaluations were done by a neurologistwho was blinded to findings on the psychiatric andneuropsychological examinations.
Psychiatric ExaminationAfter giving informed consent, the patients were
administered a series of standardized quantitative mea-sures of mood, cognitive function, and physical impair-ment. We have previously shown that these instrumentsprovide reliable and valid measurements in this strokepopulation.7"9
The modified Present State Examination (PSE), astructured psychiatric interview that elicits symptomsrelated to depression and anxiety,10 was administeredby a psychiatrist. With the use of symptoms elicited bythe PSE, a diagnosis of major depression was madeusing DSM-III criteria.11 The method for conversion ofPSE symptoms to DSM-III diagnostic criteria and theirconcurrence with Research Diagnostic Criteria wasdemonstrated in a previous publication.12
We have shown that the use of DSM-III symptomclusters for major and dysthymic (minor) depressionmakes an important distinction between these two typesof depression. Major and minor depression have differ-ent associations with lesion location,13 response to thedexamethasone suppression test,14 dementia of depres-sion,15'17 clinical severities,12 and longitudinal course.18
The Hamilton Rating Scale for Depression,19 a 17-item interviewer-rated scale, was used to measure psy-chological and physiological symptoms of depression.The Hamilton Rating Scale for Anxiety,20 a 14-iteminterviewer-rated scale, was used to measure the sever-ity of generalized or persistent anxiety.
The Mini-Mental State Examination (MMSE)21 andthe Johns Hopkins Functioning Inventory (JHFI)8 werealso administered to each patient. The JHFI contains 10items and evaluates the degree of independence in theactivities of daily living. Scores range from 0 to 27, andhigher scores indicate greater degrees of impairment.Quantitative assessments of social functioning weremade using the Social Ties Checklist (STC), whichassesses the presence of available social supports.22
The psychiatric evaluations were carried out by apsychiatrist blinded to the rest of the neurological,neuroradiological, and neuropsychological data.
Neuropsychological ExaminationThe neuropsychological examination was carried out
by a neurologist blinded to the remaining data, using abattery of tests that assessed the following domains.
Neglect and related phenomena. Patients were exam-ined for the presence of extinction on double-simulta-neous stimulation in three sensory modalities (auditory,visual, and somatesthetic). This testing was done using10 double-simultaneous stimulations intermixed with 20single stimulations (10 on each side) in a counterbal-anced order. The subject was requested to respond bypointing with one finger to the side(s) stimulated. In thecase of hemianoptic patients, the double-simultaneousstimulation was administered within the normal visualhemifield. While some patients had hemisensory defi-cits, all patients could recognize the unilateral singlestimulation. The score was the number of errors on left,right, or bilateral stimulations. Allesthetic responses(i.e., reporting the stimulation on the side contralateralto the real stimulus) were also scored.
Hemispatial neglect. Hemispatial neglect was assessedby asking the patient to copy a complex figure (asymmetrical figure with 10 lines to the left of the midlineand 10 lines to the right) and draw five figures frommemory (flower, cube, star, clock, and house). The copyof the complex figure was scored as the total number oflines omitted, and the drawing of figures from memorywas scored as the number of figures with omissions onany major part of a lateralized subcomponent (maxi-mum score was 5). Patients with significant motordeficits in their dominant hand were not given this testbecause figures may not be adequately drawn with thenondominant hand (exclusion of these patients mayhave biased the results on this test).
Personal neglect. Patients were asked to touch thehand contralateral to the lesion with the hand ipsilateralto the lesion. Scores ranged from 0 (the patientpromptly reached for the target) to 3 (no movementtoward the target).23
Motor neglect. Patients were asked to close and openboth hands 20 times. Patients with moderate to severemotor deficits were not given this test. The maximumscore was 20.23
Motor impersistence. Patients were asked to keep theireyes closed. The score was the number of seconds theeyes remained closed. The trial was discontinued after30 seconds.
Recognition of facial emotion. Patients were shown 20Eckman and Friesen faces with sad, happy, or angryexpressions.24 The patients were asked to identify theemotion in the face, using a multiple-choice answersheet that included vertically arranged schematic draw-ings of faces with sad, happy, or angry expressions, bypointing to the correct face.
Receptive aprosody. Patients listened through head-phones to 13 different six- or seven-word sentences thatwere recorded by a female speaker using a happy (fivesentences), sad (five sentences), or angry (three sen-tences) tone of voice. All sentences were simple declar-ative statements, such as "The lamp is on the table."The order of the sentences was randomized on the tape
by guest on June 29, 2018http://stroke.ahajournals.org/
Dow
nloaded from
1448 Stroke Vol 23, No 10 October 1992
with 10-second intertrial intervals. The stimulus tapewas played on a stereophonic recorder and binaurallydelivered (through stereophonic headphones) at a com-fortable hearing level. To test for possible auditorydeficits, patients were asked to repeat the first phrase.Then, patients were asked to identify the emotionalprosody and to ignore the semantic content of thesentence. Patients were provided with a response cardthat had three vertically arranged line drawings of sad,happy, or angry faces and were asked to point to theappropriate face.
Anosognosia questionnaire. This was constructed usingquestions identified by other investigators23'25 as indic-ative of anosognosia for motor and visual deficits (seeAppendix). Based on the responses obtained on theanosognosia questionnaire, patients were classified intothe following arbitrary groups: no anosognosia, thepatient's current disorder (i.e., stroke or stroke-relatedimpairment) was spontaneously reported or mentionedafter a general question about the patient's complaint;mild anosognosia, the patient's current disorder wasreported only after a specific question about thestrength of the patient's limb or the presence of visualfield deficits; moderate anosognosia, the patient's disor-der was acknowledged only after its demonstrationthrough routine techniques of neurological examina-tion; and severe anosognosia, there was no acknowledg-ment of the disorder after asking the patient aboutspecific impairments and demonstrating the existence ofeither motor or visual field deficits.
Neuroradiological ExaminationCT scans were obtained during the first week after
the stroke from all patients included in the study andwere read blinded to the clinical findings. All CT scanswere carried out in a General Electric 9900 CT scanner(Milwaukee, Wis.), and 5-mm-thick slices were ob-tained parallel to the canthomeatal line.
The damaged area was localized in specific brainregions according to the procedure of Levine andGrek.26 The anteroposterior location of each lesion wasdefined as the mean distance of the anterior border ofthe lesion from the frontal pole, averaged over all slicesin which the lesion was visible.13 This distance wasexpressed as a percentage of the maximum distancebetween the anterior and posterior poles on the CTslice in which the lesion was identified. The posteriorborder of each lesion was calculated using the samemethods. Lesion area (expressed as a percentage of thetotal brain area) was calculated from the ratio of thelargest cross-sectional area of the lesion to the area ofthe brain slice that included the body of the lateralventricles.13 We have previously demonstrated the reli-ability of this procedure and its high correlation withother methods of determining lesion volume.15
The measurements of cortical and subcortical atrophywere16: frontal horn ratio (area of the frontal horncontralateral to the lesion at the level of the foramen ofMonro, divided by the area of the whole brain at thesame level, multiplied by 1,000), third ventricle ratio(area of the third ventricle at the level of the pinealgland, divided by the area of the whole brain at the samelevel, multiplied by 1,000), lateral ventricle ratio (areaof the lateral ventricle contralateral to the lesion at thewaist level, divided by the area of the whole brain at the
TABLE 1. Demographic Data for 80 Stroke Patients
No Mild Moderate Severeanosognosia anosognosia anosognosia anosognosia
(n=53) (n=8) (n=9) (n = 10)
Age (years) 56±11 62±11 69±13 62±11Days sincestroke 5.7±3.9 7.1 ±4.9 6.7±4.8 5.6±3.5
Values are mean±SD.
same level, multiplied by 100), and four cortical sulciratio (sum of the widths of the four widest sulci dividedby the area of the whole brain at the same level,multiplied by 100). All measurements were made usinga digitizing tablet (Bioquant, Houston Instruments,Austin, Tex.) attached to an IBM AT computer.
Statistical AnalysisStatistical analysis was carried out using means and
standard deviations, one-way analysis of variance withplanned comparisons, and t tests. Frequency distribu-tions were analyzed using contingency tables and x1
tests. Regressions were calculated using a stepwiseprocedure. All probability values are two tailed.
ResultsDemographic Findings
Of the 80 patients, eight (10%) showed mild anosog-nosia, nine (11%) moderate anosognosia, 10 (13%) se-vere anosognosia, and the remaining 53 (66%) noanosognosia. There were no significant between-groupdifferences in demographic variables such as race, sex,education, alcoholism, and presence of family and per-sonal history of psychiatric disorders (data not shown).Patients were mainly black and from a low socioeconomicclass. The patients' ages and the time from the stroke toevaluation for anosognosia are given in Table 1.
Neurological FindingsThere were no significant between-group differences
in motor or sensory deficits, cerebellar and brain stemsigns, or frequency or type of aphasia (data not shown).Patients with anosognosia showed significantly higherfrequencies of hemineglect and hemianopsia than pa-tients without anosognosia (Table 2). In the mild ano-sognosia group, all eight patients with hemiparesis hadanosognosia for the motor deficit, while none of thethree patients with hemianopsia had anosognosia forthe visual field deficit. In the moderate anosognosiagroup, all nine patients with hemiparesis had anosogno-sia for the motor deficit, while none of the four patientswith hemianopsia had anosognosia for the visual fielddeficit. Finally, in the severe anosognosia group all eightpatients with hemiparesis had anosognosia for the mo-tor deficit, while two of the five patients with visual fielddeficits (four with hemianopsia and one with corticalblindness) had anosognosia for the visual field deficit. Insummary, while all 25 patients with hemiparesis andanosognosia had anosognosia for the motor deficits,only two of the 12 patients with visual field defects hadanosognosia for the visual deficit (*2=28.5, df=l,p<0.001). While this finding demonstrates that anosog-nosia for motor deficit is significantly more frequentthan anosognosia for visual field deficit, the means of
by guest on June 29, 2018http://stroke.ahajournals.org/
Dow
nloaded from
Starkstein et al Stroke and Anosognosia 1449
TABLE 2. Results of Neurological Examination in 80 Stroke Patients
Hemineglect*LeftRight
Visual field deficitstLeft hemianopsiaRight hemianopsiaCortical blindness
Side of neurological signs$LeftRightBilateral
Noanosognosia
%
4
0
11
2
0
45
45
10
No.
2
0
6
1
0
24
24
5
Mildanosognosia
%
75
0
38
0
0
088
12
No.
6
0
3
0
0
07
1
Moderateanosognosia
%
56
0
44
0
0
3367
0
No.
5
0
4
0
0
36
0
Severeanosognosia
%
40
0
40
0
10
080
20
No.
4
0
4
0
1
0
8
2
*X2=Y1A, df=3,/?=0.0006.
t*2=22.5, df=3,p=0.0001.+*2=13.0, df=6,p<0.05.
questioning may have favored motor over sensoryreporting.
Psychiatric FindingsThere were no significant between-group differences
in terms of depression or anxiety scores, and thefrequency of depression (major and minor) was no lessfrequent among the patients with anosognosia thanamong those without (Table 3). The distribution ofmajor, minor, or no depression among the three anosog-nosia groups, however, was significantly different than arandom distribution (^2=13.9, df=6,/?<0.05). This wasthe result of a significantly higher frequency of minordepression among patients with mild anosognosia com-pared with the other groups (Table 3).
It may still be argued that patients with anosognosiawould have had a more severe depression if it were notfor anosognosia. To further examine this issue we
matched six patients with anosognosia (one with mild,one with moderate, and four with severe anosognosia)and six patients without anosognosia for lesion side(four pairs with right hemisphere lesions and two pairswith left hemisphere lesions) and location (two pairswith temporo-occipital lesions, one pair with frontaldorsolateral lesions, and three pairs with basal ganglialesions). The groups were comparable in terms of age(anosognosia group, 62.1 ±13.8 years; no anosognosiagroup, 58.5±10.4 years; paired /=0.51, difference notsignificant) and years of education (8.6±3.2 and11.9±2.6 years, respectively; paired /=1.9, differencenot significant). There were no significant differencesbetween the anosognosia and no anosognosia groups inPSE score (14.8±16.5 and 8.1 ±10.4, respectively;paired /=0.83). Moreover, three of the six patients withanosognosia were depressed (two had major and onehad minor depression) compared with two of the six
TABLE 3. Psychiatric Findings in 80 Stroke Patients
Parameter
Present State ExamHamilton DepressionHamilton AnxietyMajor depression
%
No.
Minor depression*%
No.
No depression%
No.Mini-Mental State ExamtJohns Hopkins Functioning Inventory^Social Ties Checklist
No anosognosia
9.7+12.05.1+7.25.1+8.5
17
9
11
6
72
3924.3±5.65.2+4.62.9+2.0
Mild anosognosia
12.0+8.78.8+9.91.0+1.6
13
1
50
4
37
3
22.3 ±6.59.7±6.05.0±2.0
Moderate anosognosia
12.4±14.08.3±9.76.0±6.2
22
2
0
0
78
7
16.5 ±8.610.2±4.12.8 ±1.6
Severe anosognosia
7.3 ±12.53.3 ±6.33.1 ±6.8
10
1
0
0
90
922.1 ±7.78.0±5.84.0±2.5
Values are mean±SD score unless noted.V=13.2, df=3,p=0.0042.tF3,76=3.50,p<0.02; no anosognosia vs. moderate anosognosia, p<0.002.tf3,76=4.01,p<0.01; no anosognosia vs. mild anosognosia, p<0.02; no anosognosia vs. moderate anosognosia, p<0.007.
by guest on June 29, 2018http://stroke.ahajournals.org/
Dow
nloaded from
1450 Stroke Vol 23, No 10 October 1992
TABLE 4. Neuropsychological Findings in 80
Parameter
ExtinctionTactileVisualAuditory
Motor impersistence
NeglectPersonalHemispatial
SpontaneousCopy
Allesthesia
%No.
Facial emotionReceptive aprosody
Values are mean±SD*df=3,69.tdf=3,51.tdf=3,51.
Noanosognosia
1.4±3.20.8±2.50.8±2.4
28.3 ±6.5
0.16±0.46
0.05±0.310.63 ±2.2
00
16.2±3.09.9±2.9
score unless noted.
Stroke Patients
Mildanosognosia
4.5±5.05.3±4.86.1 ±4.9
21.3±13.4
0.50±0.84
0.75 ±1.54.5±8.3
252
13.4±2.37.5±3.1
Moderateanosognosia
7.5 ±4.37.1±4.46.0±4.9
15.3±12.4
0.88 ±1.13
2.2±2.598.8±8.6
111
12.5±2.56.7±2.8
Severeanosognosia
5.9±4.46.0±3.74.8±3.7
27.2±8.3
0.88±0.99
0.8±l.l2.6±4.1
303
12.1 ±4.47.5+1.9
•^3,76
4.215.812.36.68
4.94*
9.04t7.2*—
6.322.96
X2
————
—
-—
10.4
—
-
P
0.008< 0.0001<0.0001
0.0005
0.003
<0.0001<0.0004
0.015
<0.0008<0.05
patients with no anosognosia (one had major and onehad minor depression). Thus, even after controlling forimportant demographic factors and variables previouslyshown to be relevant to the production of poststrokedepression,16 no support was found for the psychologi-cal theory of anosognosia.
On the JHFI, patients with mild or moderate anosog-nosia showed significantly more impairments than pa-tients with no anosognosia (Table 3). On the MMSE,patients with moderate anosognosia had significantlylower scores than the no anosognosia group (Table 3).Differences in JHFI and MMSE scores were also ana-lyzed using more appropriate nonparametric tests, andthe findings remained significant (Kruskal-Wallis/ /= 10.56, p=0.01 and H=9A3, p=0.02, respectively).Finally, no significant between-group differences wereobserved for scores on the STC.
Neuropsychological FindingsPatients with anosognosia (mild, moderate, or severe)
showed significantly more severe tactile, visual, andauditory extinction contralateral to the lesion on dou-ble-simultaneous stimulation than patients withoutanosognosia (Table 4). Those with anosognosia alsoshowed a significantly higher frequency of motor imper-sistence (i.e., they kept their eyes closed for a signifi-cantly shorter time) and had significantly more severepersonal and hemispatial neglect (Table 4).
On the recognition of facial emotion task, patientswith anosognosia performed significantly worse than theno anosognosia group, and similar results were found onthe receptive aprosody test (Table 4).
Neuroradiological FindingsOf the 80 patients, 66 (83%) had positive CT scans.
Among the patients with negative CT findings, six hadright hemisphere neurological signs, six had left hemi-sphere signs, and two had bilateral signs.
The first significant finding was that patients withanosognosia had a higher frequency of right hemispherelesions (Table 5). Patients with mild or severe anosog-nosia showed significantly higher frequencies of tem-poroparietal and thalamic lesions than patients with noor moderate anosognosia (Table 5). On the other hand,patients with moderate anosognosia had a significantlyhigher frequency of lesions involving the basal ganglia(Table 5). No significant between-group differenceswere observed for other lesion locations. There were nosignificant between-group differences in the frequency(*2=0.30, df=3) or side (*2=3.6, df=3) of hemorrhages(data not shown). No hemorrhage was found to displacelarge portions of the hemisphere.
Since three of the six patients who were excludedfrom the study because of severe language deficits hadleft temporoparietal lesions, the finding of a significantassociation between right temporoparietal lesions andanosognosia may be related to the exclusion of thesepatients. However, when all three patients are includedin the statistical analysis as "anosognosics," there is stilla higher frequency of anosognosia in patients with righttemporoparietal lesions (11 of 16, 69%) than in patientswith left temporoparietal lesions (three of six, 50%).
Brain atrophy was measured in 51 patients. Theremaining patients were excluded because of brainstem, cerebellar, or bilateral lesions, tilted images, orunavailability of the CT scans. Owing to the few pa-tients left in each anosognosia group, all patients withanosognosia (five with mild, eight with moderate, andfive with severe) were grouped together. Patients withanosognosia showed significantly larger frontal hornand lateral ventricle as well as third ventricle ratios(Table 6). On the other hand, there was no significantbetween-group difference in the four cortical sulci ratio(Table 6). Even though patients with anosognosia had a(nonsignificantly) larger lesion area, the significant as-sociation between atrophy and anosognosia still held upafter forcing lesion area into the regression analysis as
by guest on June 29, 2018http://stroke.ahajournals.org/
Dow
nloaded from
Starkstein et al Stroke and Anosognosia 1451
TABLE 5. Lesion Location in 66 Stroke Patients With Positive Computed Tomograms
Location
HemisphereLeft*
RighttBilateralFrontalTemporoparietaltSuperioparietalBasal ganglia§ThalamustOccipitalInsulaPosterior internal capsuleCorona radiataPosterior fossa
No
%
39
44
7
20
20
7
37
5
10
10
15
12
10
anosognosia(«=41)
No.
16
18
3
8
8
3
15
2
4
4
6
5
4
Mild
%
0
88
12
12
50
12
12
38
0
12
25
0
0
anosognosia(n=8)
No.
0
7
1
1
4
1
1
301
2
0
0
Moderate anosognosia(n=9)
%
33
67
0
11
11
11
78
0
11
22
11
11
0
No.
3
6
0
1
1
1
7
0
1
21
1
0
Severe
%
0
75
25
12
75
38
2525
38
12
25
25
0
anosognosia(n=8)
No.
0
6
2
1
6
3
2
2
3
1
2
2
0
•^ 2=4.44, df=3,/>=0.035.TA-2=10.3, df=3,/?=0.015.$*2=9.51, df=3,/>=0.023.§*2=8.87, df=3,p=0.031.
the first factor (frontal horn ratio R2=0.30, F=8.87,/><0.01; third ventricle ratio i?2=0.28, F=7.59,p<0.01;lateral ventricle ratio R2=0.22, F=4.62, p<0.05). Fi-nally, no significant between-group differences wereobserved in the distance of the anterior or posteriorborder of the lesion from the frontal pole (Table 6).
To examine the relative importance of the differentfactors significantly associated with anosognosia, wecarried out a stepwise regression analysis in whichanosognosia category (i.e., no, mild, moderate, or se-vere) was the dependent variable and extinction ondouble-simultaneous stimulation, hemispatial neglecton drawing, recognition of facial emotion, receptiveaprosody, subcortical atrophy, lesion location, and im-pairments in activities of daily living (JHFI) were theindependent variables. Two variables accounted for57% of the variance: the frontal horn ratio (R=0.69,R2=QA9, F=21.0, p<0.01) and lesion location (righttemporoparietal or thalamic) (R = 0J5, R2=0.57,F=4.03,/?<0.05).
DiscussionThe present study examined the frequency and clin-
ical correlates of anosognosia in a series of patients with
TABLE 6. Neuroradiological Findings in 51 Stroke Patients
ParameterAnterior border (%)Posterior border (%)Lesion area (%)Frontal horn ratio*Third ventricle ratiotLateral ventricle ratiotFour cortical sulci ratio
No anosognosia(«=33)
39.3 + 14.661.6±20.85.3±8.10
14.8±6.77.0±4.15.7±2.22.5±1.4
Anosognosia(« = 18)
32.2±13.663.8±20.58.9±8.83
22.0±8.510.4±5.27.7±3.72.5 ±1.4
Values are mean±SD."7=3.33, df=49,/>=0.0017.17=2.49, df=49,/>=0.016.t«=2.43, df=49,p=0.018.
acute stroke lesions who were capable of undergoing averbal interview and showed several important findings.First, patients with anosognosia had a significantlyhigher frequency of neglect and related disorders, suchas extinction on double-simultaneous stimulation, mo-tor impersistence, and allesthesia. Patients with anosog-nosia also had significantly more impairments in therecognition of facial emotions and in the comprehen-sion of emotionally intoned speech. Second, patientswith anosognosia showed a significantly higher fre-quency of right hemisphere lesions, mainly involving thesuperior temporal and inferior parietal cortex, basalganglia, and thalamus. Third, patients with anosognosiaalso showed significantly more bilateral subcortical at-rophy, as demonstrated by significantly larger frontalhorn, lateral ventricle, and third ventricle ratios. Whenpatients with anosognosia were divided into mild, mod-erate, and severe anosognosia groups, some significantbetween-group differences emerged. Patients with mod-erate anosognosia showed significantly more severecognitive deficits and a higher frequency of lesionsinvolving the basal ganglia. On the other hand, patientswith mild or severe anosognosia had higher frequenciesof lesions involving the superior temporal and inferiorparietal cortex and the thalamus. Finally, patients withmild anosognosia showed a significantly higher fre-quency of minor depression.
Before further discussion, some limitations of ourstudy should be addressed. First, because we tried toavoid unreliable evaluations, patients with severe com-prehension deficits were excluded. Because anosognosiahas been frequently reported among patients with Wer-nicke's aphasia, our finding of a significantly higherfrequency of right hemisphere lesions among patientswith anosognosia may be the result of this selection bias.Second, following Bisiach et al,23 we classified anosog-nosia into three categories (mild, moderate, and se-vere), and this classification may be somewhat arbitrary.
by guest on June 29, 2018http://stroke.ahajournals.org/
Dow
nloaded from
1452 Stroke Vol 23, No 10 October 1992
Frequency of AnosognosiaThere are very few studies that have systematically
examined the frequency of anosognosia. Cutting25 foundanosognosia in 58% of patients with right hemispherelesions (as determined by neurological examination)compared with 14% of patients with left hemispherelesions. Bisiach et al23 found a frequency of 33% formoderate or severe anosognosia among patients withright hemisphere lesions. These frequencies are similarto our rate of 34% for all patients (10% mild, 11%moderate, and 13% severe anosognosia). When onlypatients with positive CT scans were considered, thefrequency of anosognosia was 38% (12% mild, 14%moderate, and 12% severe) for all patients and 51% forthose with right hemisphere lesions.
In conclusion, anosognosia is a frequent complicationof stroke lesions and will be found in about one third ofstroke patients who can reliably respond to a verbalinterview.
Anosognosia and DepressionSeveral investigators have suggested that anosognosia
is a defense mechanism of the individual confrontedwith his own deficits, thus avoiding the emergence ofdepressive symptoms.3 If this hypothesis is correct, wewould expect to find no depression among patients withanosognosia. The present study, however, empiricallydemonstrated similar frequencies and severities of ma-jor and minor depression among patients with andwithout anosognosia. Thus, our data did not demon-strate that anosognosia "protects" stroke patients fromdepressive feelings.27
Anosognosia and Cognitive ImpairmentsAnosognosia has also been considered as secondary
to memory deficits; anosognosia may be present inpatients who cannot "remember" the existence oftheir newly acquired deficits, and the anosognosicphenomenon may constitute a confabulation.28 In sup-port of this hypothesis, our patients with moderateanosognosia accepted the presence of a neurologicaldeficit only after it was demonstrated to them throughthe neurological examination, but they denied thepresence of these deficits immediately afterward. Fur-thermore, patients with moderate anosognosia showedsignificantly more severe cognitive deficits than the noanosognosia group. On the other hand, patients withsevere anosognosia did not show significantly morecognitive deficits than the no anosognosia group. Thus,cognitive impairments are not necessary to produceanosognosia but may, in some patients, constitute apredisposing or contributing factor. Future studiesmay examine the role of memory deficits in theproduction of anosognosia using more specific mem-ory measures.
It has also been suggested that anosognosia may berelated to unawareness or inattention to the side ofspace contralateral to the lesion,2 and our data areconsistent with this hypothesis. Patients with anosogno-sia showed significantly more severe hemispatial ne-glect, personal neglect, and neglect-related phenomenasuch as extinction on double-simultaneous stimulation,motor impersistence, and allesthesia than patients with-out anosognosia. Moreover, we found a significantly
higher frequency of lesions of the right hemisphereamong patients with anosognosia, which may be relatedto the importance of the right hemisphere in attentionaland arousal mechanisms. Thus, an attentional or repre-sentational deficit of the contralateral side of the bodymay be important for the production of anosognosia. Itshould be remembered, however, that we have noevidence of a causal connection between neglect andanosognosia, and these may be independent phenom-ena both associated with right hemisphere lesions.
Anosognosia and Lesion Location
Lesion location is another important factor associatedwith the presence of anosognosia. The present studydemonstrates a relation between right thalamus andright temporoparietal cortical lesions and the existenceof anosognosia. This finding supports Nielsen's29 sug-gestion that anosognosia may be caused by thalamiclesions or isolation of the thalamus from the parietalcortex. Another finding of the present study was therelevance of subcortical lesions in the production ofanosognosia because there was a significantly higherfrequency of basal ganglia lesions among patients withmoderate anosognosia.
Anosognosia and Brain AtrophyAnother important factor associated with anosogno-
sia was the presence of subcortical atrophy. Because itwas found on the earliest CT scans taken after the onsetof symptoms of stroke (i.e., usually 2-3 days), thisbilateral subcortical atrophy probably existed before thestroke lesion. Our anosognosic patients showed signifi-cantly more severe subcortical atrophy in the area of thefrontal horns and lateral and third ventricles. Thisfinding suggests that disruption of frontal subcortical ordiencephalic structures may be important predisposingfactors for the production of anosognosia. Stuss andBenson30 proposed that regions of the frontal lobe areinvolved in self-awareness and monitoring of cognitivefunction and that anosognosia could be viewed as adeficit in self-monitoring. Goldberg and Barr31 alsoproposed deficits in internal representation or monitor-ing of output as possible mechanisms of anosognosia.Thus, the presence of frontal and diencephalic atrophymay be a necessary factor for the production of anosog-nosia after lesions in a specific area of the right hemi-sphere and may explain why not every patient with aright temporoparietal or thalamic lesion developsanosognosia.
AcknowledgmentsWe thank Drs. Lynn Speedie and Kenneth Heilman for
providing us with the aprosody tapes and for their helpfulcomments.
AppendixAnosognosia Questionnaire
1. Why are you here?2. What is the matter with you?3. Is there anything wrong with your arm or leg?4. Is there anything wrong with your eyesight?5. Is your limb weak, paralyzed, or numb?6. Hew does your limb feel?
by guest on June 29, 2018http://stroke.ahajournals.org/
Dow
nloaded from
Starkstein et al Stroke and Anosognosia 1453
If Denial Is Elicited Ask the Following:a) (arm picked up) What is this?b) Can you lift it?c) You clearly have some problem with this?d) (asked to lift both arms) Can't you see that the two arms
are not at the same level?e) (asked to identify finger movements in and out of the
abnormal visual field) Can't you see that you have aproblem with your eyesight?
Scoring0: The disorder is spontaneously reported or mentioned
following a general question about the patient'scomplaints.
1: The disorder is reported only following a specific ques-tion about the strength of the patient's limb or aboutvisual problems.
2: The disorder is acknowledged only after its demonstra-tion through routine techniques of neurologicalexamination.
3: No acknowledgment of the disorder.
References1. Babinski J: Contribution a 1'etude des troubles mentaux dans
I'hemiplegie organique cerebrale (anosognosie). Rev Neurol (Paris)1914;27:845-848
2. Heilman KM: Anosognosia: Possible neuropsychological mecha-nisms, in Prigatano GP, Schacter DL (eds): Awareness of DeficitAfter Brain Injury. New York, Oxford University Press Inc, 1991,pp 53-62
3. Weinstein EA, Kahn RL: Denial of Illness: Symbolic and Psycho-logical Aspects. Springfield, 111, Charles C Thomas Publishing, 1955
4. McGlynn SM, Schacter DL: Unawareness of deficits in neuropsy-chological syndromes. / Clin Exp Neuropsychol 1989;ll:143-205
5. De Renzi E, Faglioni P: Development of a shortened version of theToken Test. Cortex 1978;14:41-49
6. Kunitz SC, Gross CR, Heyman A, Kase CS, Mohr JP, Price TR,Wolf PA: The pilot Stroke Data Bank: Definition, design and data.Stroke 1984;15:740-746
7. Robinson RG, Benson DF: Depression in aphasic patients: Fre-quency, severity and clinical-pathological correlations. Brain Lang1980; 14:282-291
8. Robinson RG, Szetela B: Mood changes following left hemispherebrain injury. Ann Neurol 1981;9:447-453
9. Robinson RG, Starr LB, Kubos KL, Rao K, Price TR: A two-yearlongitudinal study of poststroke mood disorders: Findings duringthe initial evaluation. Stroke 1983;14:736-741
10. Wing JK, Cooper E, Sartorius N: Measurements and Classificationof Psychiatric Symptoms. Cambridge, Cambridge University Press,1974
11. American Psychiatric Association: Diagnostic and Statistical Man-ual of Mental Disorders, ed 3. Washington, DC, American Psychi-atric Press, 1980
12. Robinson RG, Kubos KL, Starr LB, Rao K, Price TR: Moodchanges in stroke patients: Relationship to lesion location. CompPsychiatry 1983;24:555-566
13. Starkstein SE, Robinson RG, Price TR: Comparison of corticaland subcortical lesions in the production of post-stroke mooddisorders. Brain 1987;110:1045-1059
14. Lipsey JR, Robinson RG, Pearlson GD, Rao K, Price TR: Dexa-methasone suppression test and mood following strokes. Am JPsychiatry 1985;142:318-323
15. Robinson RG, Bolla-Wilson K, Kaplan E, Lipsey JR, Price TR:Depression influences intellectual impairment in stroke patients.Br J Psychiatry 1986;148:541-547
16. Starkstein SE, Robinson RG, Price TR: Comparison of patientswith and without post-stroke major depression matched for sizeand location of lesion. Arch Gen Psychiatry 1988;45:247-252
17. Bolla-Wilson K, Robinson RG, Starkstein SE, Boston JD, PriceTR: Lateralization of dementia of depression in stroke patients.Am J Psychiatry 1989;146:627-634
18. Robinson RG, Bolduc PL, Price TR: Two-year longitudinal studyof poststroke mood disorders: Diagnosis and outcome at one andtwo years. Stroke 1987;18:837-843
19. Hamilton MA: A rating scale for depression. J Neurol NeurosurgPsychiatry 1960;23:56-62
20. Hamilton MA: The assessment of anxiety states by rating. BrJ MedPsychol 1959;32:50-55
21. Folstein MF, Folstein SE, McHugh PR: Mini-Mental State: Apractical method for grading the cognitive state of patients for theclinician. J Psychiatr Res 1975;12:189-198
22. Starr LB, Robinson RG, Price TR: Reliability, validity and clinicalutility of the Social Functioning Exam in the assessment of strokepatients. Exp Aging Res 1983;9:101-110
23. Bisiach E, Vallar G, Perani D, Papagno C, Berti A: Unawarenessof disease following lesions of the right hemisphere: Anosognosiafor hemiplegia and anosognosia for hemianopia. Neuropsychologia1986;24:471-482
24. Ekman P, Friesen W, Ellsworth T: Emotion in the Human Face.New York, Pergamon Press Inc, 1972
25. Cutting J: Study of anosognosia. J Neurol Neurosurg Psychiatry1978;41:548-565
26. Levine DN, Grek A: The anatomic bases of delusions after rightcerebral infarction. Neurology 1984;34:577-582
27. Starkstein SE, Berthier ML, Fedorotf P, Price TR, Robinson RG:Anosognosia and major depression in 2 patients with cerebrovas-cular lesions. Neurology 1990;40:1380-1382
28. Johnson MK: Reality monitoring: Evidence from confabulation inorganic brain disease patients, in Prigatano GP, Schacter DL (eds):Awareness of Deficit After Brain Injury. New York, Oxford Univer-sity Press Inc, 1991, pp 176-197
29. Nielsen JM: Disturbances of the body scheme: Their psychologicmechanism. Bull Los Angeles Neurol Soc 1938;3:127-135
30. Stuss DT, Benson DF: The Frontal Lobes. New York, Raven PressPublishers, 1986
31. Goldberg E, Barr WB: Three possible mechanisms of unawarenessof deficit, in Prigatano GP, Schacter DL (eds): Awareness of DeficitAfter Brain Injury. New York, Oxford University Press Inc, 1991,pp 152-175
by guest on June 29, 2018http://stroke.ahajournals.org/
Dow
nloaded from
S E Starkstein, J P Fedoroff, T R Price, R Leiguarda and R G RobinsonAnosognosia in patients with cerebrovascular lesions. A study of causative factors.
Print ISSN: 0039-2499. Online ISSN: 1524-4628 Copyright © 1992 American Heart Association, Inc. All rights reserved.
is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Stroke doi: 10.1161/01.STR.23.10.1446
1992;23:1446-1453Stroke.
http://stroke.ahajournals.org/content/23/10/1446World Wide Web at:
The online version of this article, along with updated information and services, is located on the
http://stroke.ahajournals.org//subscriptions/
is online at: Stroke Information about subscribing to Subscriptions:
http://www.lww.com/reprints Information about reprints can be found online at: Reprints:
document. Permissions and Rights Question and Answer available in the
Permissions in the middle column of the Web page under Services. Further information about this process isOnce the online version of the published article for which permission is being requested is located, click Request
can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Stroke Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:
by guest on June 29, 2018http://stroke.ahajournals.org/
Dow
nloaded from
