Depression After Stroke: The Importance of Cerebral Hemisphere Asymmetries Sergio E. Starkstein, M.D. Joseph B. Bryer, M.D. Marcelo L. Berthier, M.D. Bruce Cohen, M.D. Thomas R. Price, M.D. Robert G. Robinson, M.D.

The pattern of brain asymmetries was visualized on computed tomography (CT) scan in patients with a single acute cerebrovascular lesion. Patients were divided into those with typical or reversed frontal and/or occipital asymmetries. Among patients with a typical occipital asymmetry, those with left frontal or left basal ganglia lesions showed a significantly hi gher frequency of major depression and significantly higher depression scores than patients with similar lesion location but with reversed occipital asymmetry or those with a typical asymmetry and lesions in other (left or right) brain areas. Among patients with a reversed occipital asymmetry, there was no significant association between left frontal or left basal ganglia lesions and depression. This study demonstrates that the previously reported significant association between post-stroke major depression and lesion location is restricted to patients with a typical occipital asymmetry and is not present in patients with a reversed occipital asymmetry. (The Journal Neurosciences

of Neuropsychiatry 1991; 3:276-285)

and

Clinical

D

epression

is

a

frequent

complication

of

suggested

that

several

factors,

such

as the

use

of stan-

dardized psychiatric interviews, presence of previous brain lesions, intrahemispheric lesion location, time since Received 1, 1990.

June From

20, 1990; revised the Department

September of Psychiatry

24,

1990; accepted and Behavioral

and Neurosciences, the Johns Hopkins University Baltimore; the Department of Psychiatry, University Medicine, Iowa City; the Institute of Neurological Aires;

the

University

of Buenos

Aires,

Buenos

of Iowa

College

Copyright

of Medicine, © 1991

Iowa

American

VOLUME

City,

Psychiatric

3

October Sciences

School of Medicine, of Iowa College of Research, Buenos

Aires,

Argentina;

Department of Neurology, University of Maryland, Baltimore. reprint requests to Dr. Robinson, Department of Psychiatry,

276

stroke.

About 30% to 50% of hospitalized patients with acute stroke lesions will show either a severe (major) depression or a moderate but more persistent (minor) depression.’ While there is general agreement about the prevalence of depression among hospitalized patients with acute stroke lesions, there is less agreement about the strength of the association between post-stroke depression (PSD) and lesion location. In several studies that have included different populations of stroke patients, we showed a significantly higher frequency of major depression in patients with left frontal or left basal ganglia lesions as compared to patients with lesions in any other location.2 We also showed a significant correlation between the distance of the anterior border of the lesion from the frontal pole and depression scores (i.e., more anterior lesions were significantly correlated with higher depression scores) in patients with left hemisphere lesions but not in patients with right hemisphere lesions.2 Although some of these findings have been replicated by other authors,4 other investigators only partially replicated our findings.5 We have

and

the

Address University

IA 52242. Press,

#{149} NUMBER

Inc.

3

#{149} SUMMER

1991

STARKSTEIN

stroke, or demographic factors, may explain these discrepancies.6 We have also identified several risk factors for the production of PSD. For example, among patients with left hemisphere lesions, those with major depression had significantly more subcortical atrophy (but not cortical atrophy) than nondepressed patients matched for age, location, and size of lesion.7 In addition, among patients with right hemisphere lesions, we found that those with major PSD had a significantly higher frequency of family history of psychiatric disorders than patients with right hemisphere lesions but no depression.8 Thus, both subcortical atrophy and genetic vulnerability, as well as the other factors previously mentioned, should be taken into account when looking for an association between PSD and side of lesion. Although we previously found that the relationship between PSD and lesion location was not significantly different among left-handed (as compared to righthanded)

patients,9

one

of

the

potentially

confounding

variables that has never been systematically examined is the presence of typical or reversed brain asymmetries.’#{176}’8 In the present study, we examined the frequency and correlates of PSD among a consecutive series of acute stroke patients, with either typical or reversed frontal and/or occipital asymmetries. We hypothesized that this structural asymmetry would be related to the frequency of depression associated with left frontal and left basal ganglia infarcts.

METHODS Subjects Patients included in this study were selected from 224 consecutive admissions to the University of Maryland Hospital for treatment of acute thromboembolic infarction or intracerebral hemorrhage. About one-third of the patients had to be excluded from the study because of decreased consciousness (stupor or coma) or severe comprehension deficits (as determined by an independently administered standardized neurological examination). Of the remaining 151 than 2 months post-stroke monstrable

by computed

patients, and

80 patients had a single

tomography

(CT)

were lesion scan,

less dewhich

was compatible with their acute neurological symptoms. Patients with stroke lesions in the brainstem or cerebellum (n=14) and patients whose CT scans were tilted in the lateral plane or who had lesions producing mass effect or midline shifts (n=10) were excluded from the study. All assessments were made without knowledge of the clinical data. A final population of 56 patients met all the

inclusion

JOURNAL

criteria

OF

and

NEUROPSYCHIATRY

were

entered

into

the

study.

et a!.

Neurological and Psychiatric Examination The neurological examination and diagnosis were carried out using the criteria established by the Stroke Data Bank of the National Institute of Neurological and Communicative Disorders and Stroke.’9 Handedness was assessed by asking the patient which hand they used to write, which eye they used to aim while shooting, and which leg they used to kick a ball. Patients were considered “right-handers” when the right side was dominant for all three activities, and the remaining patients were considered

“non-right-handers.”

After giving tered a series mood,

informed consent, patients of standardized quantitative

cognitive

function,

and

physical

were adminismeasures of impairment.

We

have previously shown that these instruments give reliable and valid measurements in this stroke population.2#{176} Examinations were administered in a private room between 11:00 AM. and 2:00 P.M. to minimize any possible effect of diurnal mood variation. Clinical examinations utilized the following instruments: Modified Present State Exam (PSE)21: This is a semistructured psychiatric interview that elicits symptoms related to depression and anxiety, scored by the examiner.2#{176} Possible scores may range from 0 to 144 points. Hamilton 17-item

Rating Scale for Depression (HDS)22: This is a questionnaire measuring psychological and physiological symptoms of depression, filled out by the interviewer. Scores range from 0 to 52 points. Mini-Mental State Examination (MMSE)23: This is an 11item examination that has been found to be reliable and valid in assessing a limited range of cognitive functions in stroke patients. Scores range from 0 to 30, with a score of 23 or below indicating cognitive impairment. Hopkins Functioning Inventory (JHFI)24: The JHFI contains 10 items and evaluates the degree of independence in activities of daily living, such as walking, dressing, and eating. Scores range from 0 to 27, with higher scores indicating a greater degree of impairment.

Johns

Social Ties Checklist (STC)25: The STC quantitates the number of social connections. Scores range from 0 to 10, and higher scores indicate poorer social functioning. Using the symptoms elicited by the PSE, a depression diagnosis was made using DSM-III symptom criteria for major depression or minor (dysthymic) depression.26 Although these post-stroke mood disorders do not always meet the exact criteria for time and exclusion of organic factors

for

DSM-III

use of symptom

diagnosis,

clusters

we

for major

have

and

shown

minor

that

the

depression

277

DEPRESSION

AFTER

STROKE

allows for an important distinction between these two types of depression. Major and minor depression have a differential association with lesion Iocation,2’38 response to dexamethasone suppression test,27 cognitive impairments,u. clinical severity,20 and longitudinal course.3#{176} (See Starkstein and Robinson’ for a comprehensive review.) The method used to convert PSE symptoms to DSM-III major or minor depression diagnosis is available upon request. CT Scan Examination CT scans were read by one of the authors (S.E.S.), who was blind to the clinical findings. All the CT scans were performed with a General Electric 9900 scanner, and 10-mm thick slices parallel to the canthomeatal line were obtained. Brain asymmetries were measured following the procedure of Dewan et al.3’ and Luchins and Meltzer.32 Measurements were carried out in the lowest cut showing both the frontal horns and the trigone. The inner table of the skull, the indentation of the falx in the frontal and occipital bones, the interhemispheric fissure, the septum, and the pineal gland were traced, and the hemispheres were divided by a line drawn through these structures. In all of our CT scans, the attachment of the falx and the location of the torcula Herophili has always been in the midline, so that there were no discrepancies between the anatomic and the traced midline. Following the procedure of Andreasen et al.,33 perpendiculars were then drawn from the midsagittal line to the inner table of the skull on the right and the left, at 0.90 (frontal) and 0.16 (occipital) (left and

of the anteroposterior right) anterior (frontal)

diameter, areas, and

(occipital) areas. The Asymmetry Index lated using the following formula32:

Right Right

Area Area

-

+

Left Area Left Area

-

leaving two two posterior

(Al) was

calcu-

Al

-

Areas were planimetrically measured, using a Hipad Digitizing tablet and Bioquant (Houston Instruments; Austin, Texas) software connected to an IBM AT computer. The interrater reliability (S.E.S. and M.L.B.) was r=0.99 (n=20) for both frontal and occipital area measurements. A typical or reversed asymmetry was calculated using standardized procedures described in previous studies.3’33 The mean±standard error (SE) Al for all 56 patients was calculated, and the asymmetry was a reversed one when the Al was more than 2 SEs above the mean (occipital asymmetry [mean±2]: -1.01±1.2; frontal asymmetry [mean±2]: 1.20±0.9). Lesion area was calculated from the ratio of the largest

278

cross-sectional area of the lesion on any CT scan slice to the largest cross-sectional area of the whole brain on the slice passing through the largest cross-sectional area of the lateral ventricles.8 We previously demonstrated the reliability

of this

procedure

and

its high

correlation

with

other methods for determining lesion volume.28 Finally, the lesion was localized in specific brain areas following the procedure of Levine and Grek.35 Since we previously showed that patients with left frontal or basal ganglia lesions have a significantly higher frequency of depression than patients with lesions in any other location,3 patients in the present study were divided into those with frontal lobe and/or basal ganglia involvement (“anterior group”) and those with lesions in any other brain area (“posterior group”). Statistical Analysis Statistical analysis was and three- and two-way with

planned

tions were whenever

t-test

analyzed cell sizes

performed analysis comparisons.

using means and SDs, of variance (ANOVA) Frequency

distribu-

using x2tests, and Fisher’s exact test were O.O5). No significant differences were found in the frequency of minor depression (four cases [44%] vs. one case [20%], respectively). When patients were divided into those with anterior and those with posterior lesions, none of the patients with left lesions (anterior or posterior) showed major depression, while major depression was present in two patients (40%) with a right anterior lesion and in one patient (33%) with a right posterior lesion. Frequencies for minor depression were also not significantly different (left anterior: one case [25%], left posterior: three cases [60%], right anterior: one case [20%], right posterior: none). Finally, the hypothesis of unequal frequency of major depression in patients with left anterior lesions based on the presence of a typical (seven of eight [88%]) or a reversed occipital asymmetry (none of four) was statistically substantiated (Fisher’s exact test, p