Characteristics and correlates of poststroke depression: An Indian study.

BACKGROUND: Post stroke depression adversely affects long term outcome of stroke and increases mortality risk. Few studies have looked into the comprehensive picture of post stroke depression in past. AIM: The current study aimed to look into the phenomenology, characteristic features and various correlates of post stroke depression.
METHOD: 142 consecutive stroke patients aged 60 years or above, fulfilling the inclusion criteria were assessed. Sociodemographic and clinical data were gathered using a specially designed pro-forma. Depression, apathy and psychosis were assessed by Post stroke depression rating scale, Apathy Evaluation Scale, and Brief Psychiatric Rating Scale respectively. Groups (with or without major depression) were compared using Mann-Whitney U, chi square or Fisher's exact test. One way ANOVA was conducted to see the relations of lesion location and laterality with various clinical parameters. Kaplan-Meier survival analysis was done to see the time to develop depression. The effect sizes were reported as r and partial eta squared.
RESULTS: Guilt was significantly higher (p<.05) with lesions in parietal lobe and remaining of middle cerebral artery territory. Catastrophic reaction (p<.05) and emotional dyscontrol (p<.05) were higher for diffuse lesions, periventricular lesions and lesions in frontal/occipital lobe. BPRS score, but not apathy, had a significant positive correlation with depression (Pearson's r=.692). Mean time to develop depression after stroke was 28.34 (95% CI 22.37 to 34.31) months.
CONCLUSIONS: Post stroke depression consists of various clinically important sub-components whose occurrence varies with different lesion locations. Post stroke depression is discriminable from apathy but is related to psychosis. Copyright:

INTRODUCTION

Stroke is a devastating neurological disorder with both neurological and psychiatric complications. From aphasia, hemiparesis, or other neurodeficits to depression, psychosis, or apathy, burden of stroke can be quite heavy. Poststroke depression especially worsens the overall outcome of stroke. Patients with poststroke depression show far less recovery from functional impairments compared to nondepressed patients with stroke and are 3.4 times more likely to die during the first 10 years after stroke.[1] The role of depression in poststroke mortality has been re-emphasized in a meta-analysis.[2] Opposing theories exist to explain the relation between stroke and depression. Some propose a biological mechanism behind depression, according to which vascular insults directly affect neural circuits regulating mood.[34] Inflammatory process, genetic and epigenetic variations, white matter disease, cerebrovascular dysregulation, altered neuroplasticity, and changes in glutamate transmission have all been thought to be giving rise to various psychiatric disorders after stroke.[5] Others propose psychosocial mechanism which takes the social and psychological stressors associated with stroke as the causes behind depression.[67] However, current understanding endorses a bidirectional relation between physical and psychiatric illnesses.[89] There has been controversy regarding phenomenology of poststroke depression as well. Catastrophic reactions, hyperemotionality, and diurnal mood variations have been proposed to be more consistently associated with poststroke depression compared to endogenous depression.[5] However, as these features are widely seen in other neurological conditions without depression, question remains about their specificity for poststroke depression.[10] Researches have been carried out to find an association between specific location of stroke and subsequent development of depression. Robinson and Szetela[11] in their landmark research had shown a significant inverse correlation between the severity of poststroke depression and the distance of the anterior border of the lesion from the frontal pole. Since then, there have been a number of researches looking for an association of specific lesion location and depression though reaching a consensus has not yet been possible.

At this juncture, a study of poststroke depression, looking for the phenomenology, and the clinical and anatomical correlates was felt necessary. There have been very few studies in this field from South Asian countries, and the single study from India[12] had a small sample size and had assessed depression through an instrument which was not specific for poststroke depression. The current study attempted to overcome these methodological flaws and present a holistic picture of poststroke depression.

MATERIALS AND METHODS

Participants

This was a cross-sectional study carried out in the outpatient clinic of the department of neurology in a tertiary hospital. The hospital is situated in the southern part of Kolkata, India, with a wide catchment area. Ethical clearance was obtained from the institutional ethical committee before conducting the study. The sample consisted of 142 consecutive patients, giving written informed consent, diagnosed with stroke, and fulfilling the study criteria. Inclusion criteria included patients diagnosed with stroke by qualified neurologist(s) and patients aged over 60 years. Patients with a past history of psychiatric disorder, substance use (except nicotine), family history of psychiatric disorder, poor cognitive function so as to lead to unreliable psychiatric assessment (Hindi Mental Status Examination score ≤23), and history of stressful life events (excepting current illness) since the onset of stroke or in the past 6 months (whichever is earlier) were excluded from the study. Since the study was conducted in the outpatient clinic, all those stroke patients who were admitted to the hospital were automatically excluded.

Assessment

All the assessments were made by a single clinician. The assessments were done before looking into neuroimaging or other investigation findings. Sociodemographic and clinical data were gathered using a pro forma specially designed for the study which included all the relevant parameters as described in literature including age, sex, marital status, family type, monthly income, education, residence, current and past employment status, and past medical history among others. The details of stroke, which were filled up after the clinical assessments, included location of lesion, number, type (ischemia/hemorrhage), and the exact date of the event.

Poststroke depression was assessed by the Poststroke Depression Rating Scale (PSDRS).[13] It consists of 10 sections. The first nine sections include depressed mood, guilt feelings, suicidality, vegetative (sleep and appetite) disorders, apathy/abulia/indifference, anxiety, catastrophic reactions, difficulty in emotional control, and anhedonia. The last section detects diurnal mood variation. Each section has a score range of 0–5 with a higher score meaning worse psychopathology. The assessment consists of a semi-structured interview that follows for each section. In determining the total score, all the scores of the sections are summed up excluding the last section which does not denote depression severity.[14] Hence, the PSDRS score ranges from 0 to 45. This scale is specifically designed to detect poststroke depression, emphasizing on those symptoms such as catastrophic reaction and emotional dyscontrol which are common in poststroke depression but not in functional depression. Besides, it is the only scale to provide a holistic picture of poststroke depression, beyond just depressive symptoms.

Apathy was separately assessed by the clinician-rated version of the Apathy Evaluation Scale.[15] It is an 18-item scale with each item generating a score from 1 (“not at all characteristic”) to 4 (“a lot characteristic) on a Likert scale. The maximum total score is thus 72. This is a widely used scale for assessing apathy, especially in patients with neurological illnesses.

For screening of cognitive impairment, the Hindi Mental Status Examination[16] was used as all the participants in the study could speak and understand Hindi. This is a modified version of the Mini Mental Status Examination (MMSE).[17] In this scale the main advantage over MMSE is that it can be used in illiterate persons as well. The maximum score is 30, and the cutoff for cognitive dysfunction is 23.

To screen stressful live events, the Presumptive Stressful Life Events Scale[18] was used which is an Indian adaptation of the Social Readjustment Rating Questionnaire.[19] A total of 51 life events are included in the scale which are subclassified according to whether they are personal or impersonal (not dependent on the individual's action) and according to whether they are desirable, undesirable, or ambiguous.

Psychosis was assessed by the Brief Psychiatric Rating Scale (BPRS).[20] The Global Assessment of Functioning (GAF)[21] scale was employed to assess patients' functioning in the past 1 month. Patients were also assessed by the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV)[22] to detect whether or not they had major depression.

Statistical analysis

The collected data were statistically analyzed using the Statistical Package for the Social Sciences (SPSS) 16.0 for Windows (SPSS Inc., Chicago, Ill, USA). One-way analysis of variance (ANOVA) was conducted to see the relations of lesion laterality (left-hemispheric lesion, right-hemispheric lesion, and bilateral lesion) as well as lesion location with various clinical profiles including apathy, scores in the subscales of PSDRS, BPRS score, and GAF score. Lesion location was divided into four subgroups according to predominant source of blood supply: Area A (74 patients): basal ganglia (supplied mainly by the middle cerebral artery [MCA]); Area B (28 patients): parietal lobe and rest of the MCA territory; Area C (23 patients): brain stem, cerebellum, and thalamus (supplied mainly by the posterior cerebral, vertebral, and basilar arteries); and Area D (17 patients): the remaining locations including diffuse lesions, periventricular region, frontal lobe, occipital lobe, and other locations.

According to the DSM-IV, patients were grouped as with or without major depression. The groups were compared using the Mann–Whitney U-test for continuous variables and the Chi-square test or Fisher's exact test (whichever is applicable) for categorical variables. Finally, Kaplan–Meier survival analysis was done to see the time to develop depression after stroke. The effect sizes were reported as r (calculated from z value of the Mann–Whitney U-test) and partial eta squared. The level of significance was set at P < 0.05 (two-tailed).

RESULTS

Group comparison

As shown in Table 1, patients with major depression had lower mean age, years of education, GAF score, and apathy and higher mean time since stroke, but the differences did not reach statistical significance. However, depressed patients had a significantly higher BPRS score (p< 0.001; r = 0.34) compared to their nondepressed counterparts.

Table 1

Group differences in clinical and demographic profile

	With major depression (n=55)	Without major depression (n=87)	P
Age, mean (SD)	67.42 (7.073)	68.02 (7.251)	0.591
Years of education, mean (SD)	9.16 (5.711)	9.71 (5.521)	0.619
Time since stroke (months), mean (SD)	14.75 (10.261)	13.94 (8.677)	0.766
GAF, mean (SD)	71.42 (6.682)	74.74 (6.355)	0.007
AES (C), mean (SD)	34.11 (8.728)	34.89 (9.257)	0.753
BPRS, mean (SD)	15.91 (5.093)	12.47 (4.948)	0.000*
Sex (male), n (%)	30 (54.54)	48 (55.17)	0.942
Marital status (married), n (%)	51 (92.73)	83 (95.40)	0.501
Family type (nonnuclear), n (%)	45 (81.82)	59 (67.82)	0.066
Lesion location, n (%)
Basal ganglia	26 (47.27)	48 (55.17)	0.626
Parietal lobe + MCA territory	13 (23.64)	15 (17.24)
Brain stem + cerebellum + thalamus	8 (14.54)	15 (17.24)
Others	8 (14.54)	9 (10.34)
Lesion laterality (n=141), n (%)
Left	23 (42.59)	44 (50.57)	0.144
Right	18 (33.33)	33 (37.93)
Bilateral	13 (24.07)	10 (11.49)
Current lesion number, n (%)
1	40 (72.73)	70 (80.46)	0.334†
2	3 (5.45)	6 (6.89)
Multiple	12 (21.82)	11 (12.64)
Lesion type (ischemia), n (%)	40 (72.73)	68 (78.16)	0.460
Religion (Hindu), n (%)	41 (74.54)	72 (82.76)	0.237
Income (middle), n (%)	32 (58.18)	56 (64.37)	0.459
Occupation, n (%)
Employed	4 (7.27)	5 (5.75)	0.961†
Unemployed	19 (34.54)	30 (34.48)
Others	32 (58.18)	52 (59.77)
Residence, n (%)
Urban	9 (16.36)	16 (18.39)	0.325
Rural	33 (60.00)	59 (67.82)
Semi-urban	13 (23.64)	12 (13.79)
Family diagnosis (n=40), n (%)
Medical	12 (75.00)	15 (62.50)	0.408
Stroke/related	4 (25.00)	9 (37.50)
Family history (absent), n (%)	39 (70.91)	63 (72.41)	0.846
Past medical history (absent), n (%)	43 (78.18)	68 (78.16)	0.998

†Fischer’s exact test conducted for these variables; *P<0.001. MCA – Middle cerebral artery; AES (C) – Apathy Evaluation Scale (clinician version); GAF – Global Assessment of Functioning Scale; BPRS – Brief Psychiatric Rating Scale; SD – Standard deviation

In the distribution of sex, marital status, family type (nuclear vs. nonnuclear), lesion location, laterality (right/left hemispheric or bilateral), number (one/two/multiple) and type (ischemia vs. hemorrhage), religion (Hindu vs. Muslim and others), income (middle vs. low), occupation (employed, unemployed, and others including students and homemakers), residence (urban, rural, and semi-urban), family history (present vs. absent), past medical history (present vs. absent), and family diagnoses (medical vs. stroke/related history), patients with major depression were comparable to those without.

Relations of lesion laterality with other clinical parameters

One-way ANOVA [Table 2] between various clinical parameters and lesion laterality did not show left-hemispheric lesions to be associated with a higher value of guilt in PSDRS, but the association was not statistically significant. Likewise, for the subscale of anxiety in the PSDRS, left-hemispheric lesions showed a higher value compared to right-hemispheric or bilateral lesion, but the difference failed to reach statistical significance.

Table 2

Relation of lesion laterality with various clinical profiles

	Laterality, mean (SD)			ANOVA
	Left (n=67)	Right (n=51)	Bilateral (n=23)	F	P
PSDRS total	18.10 (5.745)	16.90 (5.766)	16.39 (4.755)	1.102	0.335
PSDRS mood	2.12 (0.616)	1.98 (0.616)	1.91 (0.733)	1.200	0.304
PSDRS guilt	1.48 (0.841)	1.12 (0.739)	1.39 (0.839)	2.694	0.055
PSDRS suicidality	0.99 (0.639)	0.98 (0.678)	1.09 (0.668)	0.241	0.787
PSDRS sleep	1.43 (0.633)	1.31 (0.616)	1.35 (0.647)	0.546	0.580
PSDRS appetite	1.04 (0.614)	0.96 (0.599)	0.91 (0.417)	0.562	0.571
PSDRS apathy-indifference	2.76 (1.615)	2.75 (1.468)	2.83 (1.497)	0.022	0.978
PSDRS anxiety	2.34 (1.175)	1.94 (1.333)	1.74 (0.810)	2.928	0.057
PSDRS catastrophic reaction	1.04 (0.638)	1.06 (0.858)	0.78 (0.422)	1.419	0.245
PSDRS emotional dyscontrol	1.15 (0.530)	1.16 (0.612)	1.04 (0.475)	0.377	0.687
PSDRS anhedonia	3.75 (1.119)	3.65 (1.110)	3.35 (1.613)	0.931	0.397
PSDRS diurnal variation	1.78 (0.850)	1.69 (0.860)	1.87 (0.757)	0.405	0.668
GAF	73.67 (6.609)	73.94 (6.373)	72.52 (6.625)	0.387	0.680
AES (C)	35.18 (9.326)	33.35 (8.054)	35.65 (10.421)	0.772	0.464
BPRS	14.24 (5.027)	13.10 (5.711)	13.78 (4.843)	0.682	0.507
Time since stroke (months)	14.57 (11.187)	13.04 (7.403)	15.74 (6.635)	0.722	0.488

PSDRS – Poststroke Depression Rating Scale; AES (C) – Apathy Evaluation Scale (clinician version); GAF – Global Assessment of Functioning Scale; BPRS – Brief Psychiatric Rating Scale; ANOVA – Analysis of variance; SD – Standard deviation

Relations of lesion location with other clinical parameters

One-way ANOVA [Table 3] showed significant group differences in apathy (P = 0.036; df: 3, 138) where patients with lesions in Area C (brain stem, cerebellum, and thalamus) had the highest value among different lesion locations. For the subscale of guilt in the PSDRS, Area B, i.e., parietal lobe and the remaining MCA territory lesions, produced a significantly higher (P = 0.025; df: 3, 138) score compared to other locations. BPRS score (P = 0.001; df: 3, 138; partial eta squared = 0.112), catastrophic reaction (P = 0.039; df: 3, 138), and emotional dyscontrol (P = 0.013; df: 3, 138) were significantly higher in patients with lesions in Area D (diffuse lesions, frontal lobe, occipital lobe, periventricular region, etc.) compared to other lesion location areas. Area D also registered a higher value than other areas, for total PSDRS score (P = 0.070; df: 3, 138) and the subscale of sleep (P = 0.089; df: 3, 138). The differences, however, did not reach statistical significance. Similarly, a higher mean suicidality score (another subscale of PSDRS) was seen with lesions in Area B which failed to achieve statistical significance (P = 0.057; df: 3, 138).

Table 3

Relation of lesion location with various clinical profiles

	Lesion location, mean (SD)				ANOVA
	Area A (n=74)	Area B (n=28)	Area C (n=23)	Area D (n=17)	F	P
Age (years)	68.46 (7.753)	66.79 (6.641)	67.35 (7.030)	67.12 (5.566)	0.473	0.702
Years of education	9.64 (5.877)	8.86 (5.407)	9.70 (5.764)	9.71 (4.593)	0.153	0.928
Time since stroke (months)	13.31 (8.158)	15.62 (10.171)	14.23 (12.356)	15.71 (7.389)	0.547	0.651
GAF	74.42 (6.857)	73.32 (7.318)	72.57 (5.017)	70.65 (6.133)	1.689	0.172
AES (C)	32.85 (8.551)	34.50 (10.423)	38.70 (7.945)	36.71 (8.498)	2.925	0.036*
BPRS	12.70 (4.742)	15.64 (4.490)	12.48 (5.125)	17.35 (6.623)	5.827	0.001*
PSDRS total	16.84 (5.075)	19.25 (6.108)	16.09 (6.338)	19.53 (6.663)	2.403	0.070
PSDRS mood	2.00 (0.549)	2.21 (0.833)	1.91 (0.733)	2.24 (.831)	1.410	0.242
PSDRS guilt	1.24 (0.824)	1.75 (1.076)	1.13 (0.757)	1.53 (.624)	3.214	0.025*
PSDRS suicidality	0.88 (0.618)	1.25 (0.752)	1.04 (0.638)	1.18 (.728)	2.573	0.057
PSDRS sleep	1.27 (0.556)	1.57 (0.690)	1.39 (0.656)	1.59 (.795)	2.214	0.089
PSDRS appetite	0.97 (0.596)	1.07 (0.604)	0.87 (0.548)	1.18 (.529)	1.097	0.353
PSDRS apathy	2.74 (1.481)	3.07 (1.609)	2.52 (1.620)	2.71 (1.532)	0.578	0.631
PSDRS anxiety	2.04 (1.091)	2.18 (1.188)	1.83 (1.114)	2.71 (1.687)	1.948	0.125
PSDRS catastrophic reaction	0.95 (0.639)	1.11 (0.832)	0.87 (0.626)	1.47 (1.007)	2.875	0.039*
PSDRS emotional dyscontrol	1.05 (0.402)	1.14 (0.591)	1.09 (0.515)	1.53 (.874)	3.729	0.013*
PSDRS anhedonia	3.69 (1.146)	3.89 (1.315)	3.43 (1.273)	3.41 (1.228)	0.862	0.462
PSDRS diurnal variation	1.74 (0.812)	1.64 (0.826)	1.91 (0.793)	1.76 (1.033)	0.443	0.723

*P<0.05. Area A – Basal ganglia; Area B – Parietal lobe and rest of the MCA territory; Area – Brain stem, cerebellum, and thalamus; Area D – Others (diffuse lesions, frontal lobe, occipital lobe, periventricular region, etc.). PSDRS – Poststroke Depression Rating Scale; AES (C) – Apathy Evaluation Scale (clinician version); GAF – Global Assessment of Functioning Scale; BPRS – Brief Psychiatric Rating Scale; ANOVA – Analysis of variance; SD – Standard deviation

Other findings

Poststroke depression total score was not significantly correlated with age, years of education, time since stroke, or apathy. However, a significant positive correlation (p < 0.001) was seen with BPRS (Pearson's r = 0.692). This significance was maintained (p < 0.001) even when partial correlation between them was calculated after age, years of education, time since stroke, and apathy were controlled (Pearson's r = 0.731).

The prevalence of poststroke depression (major) in the current study was 38.73% (55 patients out of 142). Kaplan–Meier survival analysis showed [Figure 1] mean time to develop depression (as diagnosed by the DSM-IV) after stroke was 28.34 (95% confidence interval: 22.37–34.31) months.

Figure 1

Kaplan–Meyer survival analysis showing time to depression after stroke. Mean survival time – 28.34 (95% confidence interval: 22.37–34.31) months

DISCUSSION

A major strength of the current study lied in elimination of rater bias. The rater was blind to investigation reports (including neuroimaging findings) or current treatment regimen as ratings were done before looking into them. Sample selection was another area of strength. All the predisposing factors for depression, apart from the stroke itself, were eliminated by excluding patients with a past history or family history of psychiatric illness and patients with current or past substance use and stressful life events (excepting the current stroke) in a relevant time period. Hence, the depression seen in the study population was biologically and/or psychosocially related to stroke only, and the confounding variables were taken care of. The sample size was robust, and the instrument for detecting depression was specific for poststroke depression only.

The highlight of the current study finding is that though the depression score (PSDRS total) was not significantly different across lesion locations and did not change significantly according to lesion laterality, its components did. Guilt, catastrophic reaction, and emotional dyscontrol significantly varied across lesion locations. Again guilt and anxiety, though not statistically significant, did vary according to lesion laterality. These findings probably justified the selection of PSDRS over the other rating scales to detect functional depression. It produces a comprehensive picture of poststroke depression beyond the depressive symptoms. It also helped in looking into the relation of symptoms more specific to poststroke depression (as compared to functional depression, e.g., catastrophic reaction), with lesion location and laterality.

Guilt and anxiety were more (though not statistically significant) in left-hemispheric lesions. Left-hemispheric lesions have been reported to be more consistently associated with depression,[23] though the importance of laterality is diluted with passage of time after stroke. By 3–6 months poststroke, proximity to the frontal lobe in both hemispheres influences the occurrence of poststroke depression.[24] Guilt also showed significantly different values across lesion locations with the highest mean value for lesions in the parietal lobe and the remaining of the MCA territory. However, relations of components of poststroke depression-like guilt with lesion laterality or location have not been reported previously.

Apathy, abulia, and indifference, as a subscale of PSDRS, did not produce significantly different values across lesion locations, but a composite measure of apathy through the Apathy Evaluation Scale showed the highest value for lesions in the thalamus, cerebellum, or brain stem. Apathy is more commonly seen in lesions of cortico-subcortical circuits[25] and basal ganglia which lead to dysfunctions of frontal subcortical system.[26] However association of apathy with thalamic lesions has also been seen.[272829]

Catastrophic reaction and emotional dyscontrol had significantly higher mean values for diffuse lesions, periventricular lesions, and lesions in the frontal and occipital lobes. This was in comparison with lesions in the parietal lobe, basal ganglia, cerebellum, brain stem, and thalamus. The current study population had a disproportionately little number of patients with lesion in the frontal and occipital lobes, and the group with lesions in Area D (the “others” group) which included these locations contained only 17 patients. Hence, this particular finding should be cautiously generalized.

Depression, as assessed by total score of PSDRS, did not have any significant correlation with apathy. Apathy has been reported to be related to but discriminable from depression.[15] The current finding thus is in accordance with that. However, poststroke depression was significantly positively correlated with psychosis (BPRS score), and the significance was maintained in partial correlation between them after the possible confounding variables were controlled. This was reflected in comparison of groups as well. Patients with depression (diagnosed by the DSM-IV) had a significantly higher BPRS score than those without. This relation between poststroke depression and psychosis has not been reported in the past. Poststroke depression and psychosis are two distinguishable entities, but BPRS might not be the best instrument to detect poststroke psychosis. Probably, this finding should encourage researchers to device an instrument to specifically detect poststroke psychosis.

The prevalence of major depression (as per the DSM-IV) was seen to be 38.73% which falls in the range described in literature.[10] Mean time to develop major depression was 28.34 months. This is longer than what is described in literature so far,[30] which reports the peak prevalence of depression to be around 3–6 months following stroke, declining by about 50% in 1 year time. However, in the current study, survival analysis was done only for major depression (as per the DSM-IV) and not minor depression or dysthymia which also are common after stroke. This might explain the longer mean time to develop depression.

CONCLUSION

It can be said that poststroke depression consists of various clinically important subcomponents apart from depressed mood, and their occurrence varies according to lesion locations. Poststroke depression is discriminable from apathy but is related to poststroke psychosis.

The current study had certain limitations. First, the cross-sectional design failed to detect the relation of stroke and depression over a longer time period. It also could not throw any light on the course and long term outcome of poststroke depression. Second, since the study was hospital based, many patients in community who were not obtaining the services of this hospital were missed. Finally, to eliminate confounding variables for depression, patients with a past or family history of depression and patients with current or past substance use were excluded from the study. This elimination was achieved against a cost as it made the current study unable to detect the association between these factors and poststroke depression. In the future, community-based prospective studies from these parts of the world might shed more light in this interesting field.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.