Cognitive Dysfunction in Stroke Survivors: A Community-Based Prospective Study from Kolkata, India

Cognitive Dysfunction in Stroke Survivors: A Community-Based Prospective Study from Kolkata, India Sujata Das, PhD,* Neelanjana Paul, MD,† Avijit Hazra, MD,‡ Malay Ghosal, MD,† Biman Kanti Ray, DM,* Tapas Kumar Banerjee, FRCP,x Prabir Burman, PhD,{ and Shyamal Kumar Das, DM*

The frequency of cognitive dysfunction among community stroke survivors (SS) is not known in India. This prospective study investigated the prevalence of poststroke mild cognitive impairment (psMCI) and poststroke dementia (psDem), the annual progression rate to dementia, and pertinent risk factors in a sample population of SS in Kolkata, India between September 2006 and July 2010. From a communitybased stroke registry, 281 SS were assessed at baseline year after excluding attrition due to various causes. Validated cognitive tools were applied by trained field workers under supervision of a neuropsychologist. The assessment was repeated annually for consecutive three years. The mean follow-up period was 1.89 years (range, 1-3 years). The period prevalence rate of psDem was 13.88% (95% confidence interval [CI], 9.91%-18.90%) at baseline, and the average annual progression rate to dementia was 3.53% (95% CI, 2.09%-5.58%). Compared with subjects without dementia, those with psDem were significantly older at first-ever stroke and more likely to have cortical atrophy. The period prevalence rate of psMCI was 6.05% (95% CI, 1.45%-13.64%) at baseline, and 10.6% (95% CI, 4.57%-20.88%) of these subjects converted to psDem annually. Survival analysis of psDem patients showed a greater risk of death in psDem SS as compared to nondemented SS (hazard ratio, 2.65; 95% CI, 1.72-6.15). Our data suggest that the overall prevalence of psDem is higher than that of psMCI, possibly related to nonexclusion of prestroke dementia, but that the average annual progression rate of SS to psDem and that of psMCI to psDem are comparable. Older age at first-ever stroke and cortical atrophy are associated with increased risk for dementia among SS. Key Words: Prevalence— poststroke dementia—poststroke mild cognitive impairment. Ó 2013 by National Stroke Association

From the *Department of Neurology, Bangur Institute of Neuroscience, Kolkata, India; †Department of Psychiatry; ‡Department of Pharmacology, Institute of Postgraduate Medical Education and Research, Kolkata, India; xDepartment of Neurology, National Neuroscience Centre, Kolkata, India; and {Department of Biostatistics, University of California, Davis, California. Received February 3, 2012; revision received March 13, 2012; accepted March 21, 2012. The study was funded by Indian Council of Medical Research, India (Project SWG/Neuro/20/2005/NCD-1). Sujata Das received remuneration as a Senior Research Fellow for this project. The other authors have no conflicts of interest to disclose. Author Contributions: Sujata Das collected data in the field, and wrote the first draft. Neelanjana Paul collected data in the field and

helped in analysis of data. Avijit Hazra undertook statistical analysis and revised the manuscript. Malay Ghosal conceptualized the project and reviewed study data. Biman Kanti Ray undertook field visits and monitored the study. Tapas Kumar Banerjee conceptualized and monitored the study. Prabir Burman undertook statistical analysis. Shyamal Kumar Das conceptualized and monitored the study, exercised overall supervision and finalized the draft. Address correspondence to Shyamal Kumar Das, Department of Neurology, Burdwan Medical College, 52/1 A, Sh7ambhu Nath Pandit Street, Burdwan, WB 713101, India. E-mail: das_sk70@ hotmail.com. 1052-3057/$ - see front matter Ó 2013 by National Stroke Association doi:10.1016/j.jstrokecerebrovasdis.2012.03.008

Journal of Stroke and Cerebrovascular Diseases, Vol. 22, No. 8 (November), 2013: pp 1233-1242

1233

S. DAS ET AL.

1234

Cognitive dysfunction (CD) is an important sequel of stroke that slows recovery, induces caregiver stress, and shortens life span.1,2 It may appear after an acute episode3-5 or during the delayed recovery phase,5-7 and can range in severity from mild deterioration of individual cognitive domain8 to dementia3,4,9 (Table 1). The reported prevalence of poststroke mild cognitive impairment (psMCI) in stroke survivors (SS) varies from 71.1% in first month to 61.3% at 6 months,8 whereas that of poststroke dementia (psDem) shows a wide variation, from 6% to 38%.3-5 The frequency of poststroke CD has been studied in community and hospital settings of many developed Western and affluent Eastern countries.3-15 In view of the aging population and increasing stroke incidence in India,16 the prevalence of poststroke CD is also expected to be high. Studies from developing countries have consistently reported higher mortality in SS with psDem than in SS without dementia.17-22 Such data are not available for developing countries, including India. In the present study, we investigated the prevalence of psMCI and psDem in an urban community cohort of SS in India. We also determined the annual progression rate (APR) of SS to psDem and the APR of psMCI to dementia, and analyzed pertinent risk factors.

Materials and Methods This community-based prospective study was conducted between September 2006 and July 2010. The study pro-

tocol was approved by Institutional Ethics Committee of the Institute of Postgraduate Medical Education and Research. Written informed consent was obtained from each study subject or a reliable family member. The prevalence and APR of psDem and psMCI among SS were determined during baseline and 3 annual follow-up visits.

Survey Area and Sampling Strategy The survey was conducted in the city of Kolkata, the largest metropolis in eastern India, with an area of 185 km2 and population of 4.58 million within municipal limits.23 The city has a heterogeneous population, with 80% of the inhabitants speaking Bengali (the regional language) and the rest mostly Hindi. The literacy rate is 81.31%.24 A stratified random sampling strategy was followed as described previously.16 In brief, the municipal area of the Kolkata is divided into 5200 smaller units known as National Sample Survey Organization (NSSO) blocks. Complete data for each block, including geographical landmarks, boundaries, and types of housing are available. Based on this information, we divided the city into 6 strata. Each stratum acted as a sampling frame. Stratum I consisted of predominantly low socioeconomic housing or slum area. Strata II-VI comprised blocks from the remaining parts of the city (central, northern, and southern). The southern and northern parts were subdivided into areas with and without consolidated housing complexes. In the central part, only consolidated housing was considered.

Table 1. Previous studies exploring cognition in SS Location and year

Type of study

Evaluated stroke sample (strength at entry)

Hong Kong, 20043 Cross-sectional, hospital-based Portugal, 20014 Cross-sectional, hospital-based London, 20025 Longitudinal, community-based

280 (484) 220 (237) 645 (1454)

Sweden, 20046 United States, 20047 The Netherlands, 20048

Cross-sectional, hospital-based Longitudinal, community-based

149 (243) 212 (321)

Longitudinal, hospital-based

196 (592)

Singapore, 20029 Spain, 200510 United Kingdom, 200211 Israel, 200212 India, 201013 Italy, 199714 Australia, 200415

Longitudinal, hospital-based Longitudinal, hospital-based Longitudinal, hospital-based

252 (270) 193 (327) 351 (360)

Longitudinal, hospital-based Cross-sectional, hospital-based Cross-sectional, community-based Cross-sectional, community-based

77 164 (164) 75 (80) 99 (458)

Outcome with time intervals Dementia at 3 months, 20% Dementia at 3 months, 6% Cognitive impairment at 3 months, 38%; 12 months, 39.25%; 36 months, 41.5%; 48 months, 41.5% Dementia at 18 months, 28% Dementia at 10 years, 19.3% Dementia at 1 month, 10.8%; 6 months, 7.7%; 12 months, 7.7%; MCI at 1 month, 71.1%; 6 months, 61.3%; 12 months, 51.5% Dementia at 6 months, 4%; 12 months, 9.03% Dementia at 3 months, 18% Dementia at 12 months, 23% Only dementia cases studied Cognitive impairment at 3 months, 31.7% Dementia at 1 month, 30% Dementia at 12 months, 12.5%; cognitive impairment no dementia at 12 months, 37.5%

COGNITIVE DYSFUNCTION AMONG STROKE SURVIVORS IN INDIA

Blocks vary in size, with an average of 75-150 households per block. From each stratum, a number of blocks were selected (proportionate to the population) using a random number list. Half of the households in each block were surveyed. A total of 282 NSSO blocks were selected, which allowed screening of a population of 100,802 for this study.

Inclusion and Exclusion Criteria For inclusion, all sampled individuals had to be residents of Kolkata for at least 1 year and to have at least 1 reliable family informant to provide relevant information. Subjects with transient ischemic attack were excluded. SS with aphasia, ongoing psychosis, visual or auditory deficits of sufficient significance to hinder retrieval of relevant and adequate information were excluded from the cohort assessed for CD.

Survey Team, Methodology, Instruments, and Follow-Up The survey team comprised 4 graduate field workers trained in epidemiologic surveys, a neuropsychologist, and a neurologist. Before the actual survey commenced, the field workers were trained to recognize stroke and cognitive impairment and to administer study tools in special clinics of our institute. The survey was conducted in 3 stages. In the first stage, a field worker administered a validated family screening questionnaire23 to the head of household or a responsible family member to detect cases of stroke. This validated questionnaire captured demographic data and contained screening questions to identify stroke, epilepsy, movement and cognitive disorders. To avoid missing cases, 2 more attempts were made to visit subjects who were not available initially. Visits were also made during evenings to ensure that housebound and unemployed individuals were not overrepresented. In the second stage, the field neurologist confirmed stroke cases and scrutinized the available clinical information. Clinical and radiological data on stroke cases were recorded in a separate case report form (stroke proforma). In the third stage, validated questionnaires for the evaluation of cognition, activities of daily living (ADL), and depression were administered by field workers under the neuropsychologist’s supervision. All data were subsequently reviewed by a team of senior neurologists and psychiatrists for consensus diagnosis. Considering a neurologist’s diagnosis as the gold standard, the screening questionnaire used to detect stroke cases had a sensitivity of 83.3% and a specificity of 99.9%.23 The validated tools used for cognitive assessment were the Bengali version of Hindi Mental State Examination (BMSE),25,26 which is an adaptation of the Mini-Mental State Examination,27 and the Kolkata Cognitive Screening Battery (KCSB).26 The BMSE was developed to suit our urban sample by appropriate modifications of items used previously in surveys of

1235 25,26

rural populations. The BMSE was used initially to screen for cognitive impairment, and scoring was done according to norms established from this population based on age, education, and sex (normative data available on request). The sensitivity and specificity of BMSE for detecting CD was 80.3% and 90%, respectively, with detection by a senior neurologist serving as the gold standard. Subjects scoring below the cutoff value of 1.5 standard deviation below the appropriate mean BMSE score were further evaluated with the KCSB.26 The cognitive functions assessed included attention, orientation, comprehension, naming, memory (immediate and delayed recall, recognition), calculation, constructional praxis, similarities, clock drawing, and category naming. The Everyday Abilities Scale for India (EASI) was used to assess ADL, such as independence in feeding, dressing, toileting, social participation, and decision making in personal and familial issues.23 The EASI was found to have a sensitivity of 80.3% and a specificity of 90%. Depressive symptoms were evaluated using the Geriatric Depression Scale (GDS), which also was validated for use in our population, with depression diagnosed by a GDS score $21.23 The severity of psDem was assessed using the Clinical Dementia Rating (CDR) scale.28 The CDR scale covers 6 domains (memory, orientation, judgment and problem solving, community affairs, home and hobbies, and personal care), rated as 0, no dementia; 0.5, questionable dementia; 1, mild dementia; 2, moderate dementia; or 3, severe dementia. This scale has been used previously in the Indian population.29 All of the available subjects were followed-up annually for 3 consecutive years after the baseline assessment. On each follow-up, information on attrition of subjects and likely causes was obtained. Cognition, ADL, and depression were reassessed at each follow-up visit by the same person using the same tools. Both the neurologist and the neuropsychologist participated in the follow-up field visits.

Operational Definitions Stroke was defined as rapidly developing clinical signs of focal (or global) disturbance of cerebral function, with symptoms lasting 24 hours or more or leading to death, with no apparent cause other than vascular origin.30 First-ever stroke (FES) was defined as stroke occurring in a patient with no previous stroke event. Imaging features of cerebral infarction and intracerebral hemorrhage have been defined previously.16 Diffuse shrinkage of cerebral cortex detected by computed tomography or magnetic resonance imaging was interpreted as cortical atrophy. Dementia was diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders, Third Edition Revised (DSM-III-R) criteria.31 Mild cognitive impairment was diagnosed based on the presence of either impairment and/or CD in any other domain

1236

with intact functional activities and absence of dementia, as in our previous study.32 psDem and psMCI were defined as dementia or MCI noted after the index stroke event.33

Statistical Analysis Analyses were performed using Statistica version 6 (Statsoft, Tulsa, OK) and SPSS version 17.0 (SPSS, Chicago, IL). The period prevalence rate (PR) was calculated as a proportion with respect to total number of stroke cases over 1 year, expressed as percentage. The prevalence of psDem after correcting for attrition was estimated using the last-observation-carried-forward method (LOCF). LOCF is an accepted strategy for dealing with missing data in which the last measurements obtained before a patient dropped out from a study involving a series of measurements are carried forward.34 The APR was calculated as number of cases in stroke cohort converting to a higher or lower level of cognitive performance each year on average, expressed as percentage. Calculation of 95% confidence intervals (CIs) for key data assumed Poisson distribution for observed cases. Numerical variables, were compared between groups by one-way analysis of variance (ANOVA) if normally distributed and by Kruskal-Wallis ANOVA if skewed. Categorical variables were compared using Fisher’s exact test or the c2 test, as appropriate. Median survival in SS with psDem, SS with psMCI, and SS without CD was determined by KaplanMeier analysis, and differences in survival were assessed using the log-rank test. Hazard ratios (HRs) were calculated. All analyses were 2-tailed, and P , .05 was considered statistically significant in univariate analysis. Risk factors returning a P value , 0.2 in univariate analysis were included in logistic regression analysis for significant risk of CD in SS. The associations were explored through both nominal (multinomial) and ordinal regression models using 3 outcome categories: no CD, psMCI, and psDem.

Results A population of 100,802 was screened. The age and sex distributions of the screened population were representative of the population of Kolkata.24 Out of a total of 853 stroke cases identified at baseline, 452 individuals died before the start of detailed assessment. Sixty-one individuals had indequate information available (mean presenting age [MPA], 63.1 6 14 years; males, n 5 36; mean age at FES, 58.1 6 16.18 years), 19 were aphasic (MPA, 66.9 6 12.64 years; males, n 5 9; mean age at FES, 61.6 6 14.25 years), 20 were absent (MPA, 61.9 6 15.31 years; males, n 5 14; mean age at FES, 61.6 6 12.31 years), and 20 refused to undergo detailed assessment (MPA, 70.7 6 10.02 years; males, n 5 10; mean age at FES, 64.7 6 9.65 years). These exclusions brought the number of those evaluated at baseline to 281. These SS were followed-up for a maximum of 3 years (mean, 1.89 years; range, 1-3 years).

S. DAS ET AL.

Figure 1 depicts the flow of 281 SS assessed in baseline year (September 2006 to August 2007) with subsequent attritions due to death, migration, absence, or refusal, and the numbers evaluated at 3 consecutive annual visits. The distribution of interval from stroke occurrence and assessment were ,1 year, 17.8%; 1-2 years, 11.7%; 2-5 years, 35.2%; and .5 years, 35.2%. For the 281 SS, the overall PR of psDem at baseline was 13.88% (95% CI, 9.91%-18.90%) (Fig 1). The frequency of psDem was 11.76% (95% CI, 1.42%-42.45%) at 3 months after FES (n 5 2/17) and 15.15% (95% CI, 4.91%-35.36%) at 6 months after FES (n 5 5/33). The estimated PR of psDem at study end, after correcting for attrition by the LOCF method, was 15.83% (95% CI, 13.31%-18.68%). The estimated mean APR for new cases of dementia from the stroke cohort was 3.53% (95% CI, 2.09%-5.58%). Out of the total 57 dementia cases recorded over the study period, 39 (68.42%) were found at baseline, and the remainder developed during follow-up. Thus, approximately 1 in 3 cases were new. Seventeen cases of psMCI were identified at baseline, and PR was calculated as 6.05% (95% CI, 3.52%-9.69%) (Fig 1). An additional 8 new cases were detected during follow-up. The 25 psMCI cases included 9 amnestic type and 16 nonamnestic type. Cognitive functioning improved in 4 SS, and 8 SS deteriorated to dementia within 3 years, for an annual conversion rate of 10.6% (95% CI, 3.45%-15.76%). Four SS with psMCI died during the study period, for an annual death rate of 5.33% (95% CI, 1.09%-10.24%). Table 2 shows the severity of CD based on CDR score. Over the assessment period, the proportion of cases of mild to moderate dementia decreased, whereas that of severe dementia increased, although the trend was not statistically significant (P 5 .247). Among the 18 SS with psDem who developed dementia during follow-up, 9 (50%) had recurrent stroke before the development of dementia and a CDR score of 3 (severe CD). The frequency of psDem in men increased with advancing age, peaking in the 70- to 79-year age group and declining thereafter (Table 3). No clear trend was evident in women. Overall, stroke recurrence was more common in SS with psDem than in SS with psMCI (49% vs 20%). The results of univariate analysis of potential risk factors for psDem and psMCI are presented in Table 4. Compared with SS without CD, SS with psDem were older at FES; sustained more recurrent strokes, cerebral infarction, and cortical atrophy; and had poorer performance on the BMSE and EASI. SS with psMCI performed worse on the BMSE, EASI, and GDS but otherwise did not differ from the SS without CD. SS with psDem were more likely than SS with psMCI to have cerebral infarction and cortical atrophy. Ordinal logistic regression analysis with all selected risk factors included returned P , .15 for the following variables: age at stroke onset (P , .001), presence of

COGNITIVE DYSFUNCTION AMONG STROKE SURVIVORS IN INDIA

1237

Figure 1. Flowchart depicting the number of SS assessed over time and the prevalence of psDem and psMCI in these SS. FU, follow-up.

cortical atrophy compared with absence (P 5 .021), male compared with female (P 5 .059) and recurrent compared with no recurrence (P 5 .135). A ‘‘best model’’ selection via stepwise regression using the Akaike information criterion identified these same risk factors with very close P values. Thus, multivariate analysis also suggested that higher age at stroke onset and cortical atrophy are important risk factors for development of CD after stroke.

However, unlike in univariate analysis, recurrence of stroke might not be a strong risk factor. In a nominal logistic model using risk factors that provide a better model, age at onset of FES and cortical atrophy clearly predicted risk of dementia compared with no CD, and age at onset predicted risk of MCI compared with no CD. However, no clear risk factor emerged to distinguish between MCI and dementia after stroke.

Table 2. Frequency and severity of CD at successive annual visits assessed through the CDR

Total stroke survivors assessed, n No CD, n* Total poststroke mild cognitive impairment cases (CDR value 0.05), n (%) Total psDem cases, n (%) Mild dementia (CDR value 1), n (%) Moderate dementia (CDR value 2), n (%) Severe dementia (CDR value 3), n (%)

First visit

Second visit

Third visit

Fourth visit

281 225 17 (6.05)

219 185 12 (5.48)

180 147 8 (4.44)

158 125 5 (3.16)

39 (13.88) 2 (5.13) 11 (28.21) 26 (66.67)

22 (10.05) 0 4 (18.18) 18 (81.82)

25 (13.89) 1 (4.00) 4 (16.00) 20 (80.00)

28 (17.72) 0 6 (21.43) 22 (78.57)

The total number of SS with psDem at some point during the study period was 57 (20.28%), and that of psMCI was 25 (8.90%). Percentages in the severity categories denote distribution within the respective columns. *SS without mild cognitive impairment or dementia.

S. DAS ET AL.

1238

– 5 7.69 10.64 6.06 7.50 6.67 6.67 6.05 54 (11) 40 (9) 39 (6) 47 (21) 33 (13) 40 (11) 15 (6) 13 (3) 281 (80)

– – 3 (1) 10 (6) 5 (2) 13 (5) 5 (4) 3 (1) 39 (19)

– – 7.69 21.28 15.15 32.50 33.33 23.08 13.88

– 2 (1) 3 5 (3) 2 (1) 3 (1) 1 1 17 (6)

Twenty-four of 57 SS with psDem died within 3 years, for an average annual mortality rate of 14.04% (95% CI, 9.08%-20.72%), compared with an annual mortality rate of 3.2% in SS without CD. A comparison of KaplanMeier plots (Fig 2) shows a significant difference in survival between SS with psDem and SS without CD (P , .001). Considering death as the endpoint, the SS with psDem were at greater risk of death, with an HR of 2.65 (95% CI, 1.72-6.15). The median survival of SS with psDem was 117.77 months. In SS without CD, median survival exceeded the duration of the study. There was no significant difference in survival between the SS with psMCI and those without CD.

22 (6) 23 (5) 20 (1) 20 (7) 15 (5) 22 (3) 8 (3) 5 (1) 135 (31)

– – 2 3 (1) 0 5 (1) 1 (1) 1 12 (3)

– – 10.00 15.00 0 22.73 12.50 20.00 8.89

– – 1 3 (2) 1 (1) 1 0 0 6 (3)

– – 5.00 15.00 6.67 4.55 0 0 4.44

Discussion

*Value in parentheses is the number of SS with recurrent stroke.

– 11.76 10.53 7.41 5.56 11.11 14.29 12.50 7.53 – 2 (1) 2 2 (1) 1 2 (1) 1 1 11 (3) – – 5.26 25.93 27.78 44.44 57.14 25.00 18.49 – – 1 (1) 7 (5) 5 (2) 8 (4) 4 (3) 2 (1) 27 (16) 32 (5) 17 (4) 19 (5) 27 (14) 18 (8) 18 (8) 7 (3) 8 (2) 146 (49) #49 50-54 55-59 60-64 65-69 70-74 75-79 801 Total

Cases of Cases of Cases of Cases of Cases of Cases of Age group, years Population* psDem* PR, % psMCI* PR, % Population* psDem* PR, % psMCI* PR, % Population* psDem* PR, % psMCI* PR, %

Overall Women Men

Table 3. Age-and sex-specific distribution of SS (n 5 281) with psDem (n 5 39) and psMCI (n 5 17) at baseline year according to age at FES

Survival Analysis

The strength of this study lies in its prospective evaluation of SS from a randomly selected community-based sample with validated tools. We found a higher prevalence of psDem than that of psMCI, but comparable annual rates of progression of SS to psDem and conversion of psMCI to psDem. Our data identify older age at FES and cortical atrophy as risk factors for psDem. A previous meta-analysis of stroke with dementia reported an overall 7.4% (95% CI, 4.8%-10.0%) rate of psDem among population-based studies.22 However, the prevalence of dementia depends on many factors, including defining criteria, methodology, mortality rate, ethnicity, demography, inclusion or exclusion of prestroke dementia, and interval between stroke occurrence and assessment. In the present study, the frequency of dementia at 3 months after FES was lower than in previous hospitalbased studies from Hong Kong3 and Spain,10 and that at 6 months after FES was higher than in similar hospital studies from The Netherlands8 and Singapore9 (Table 1). These variations could be related to admission of more severe cases in hospital setting and differences in ethnicity among the study populations. However, comparing studies of similar design, the PR of dementia in the present study was similar to that reported in a study from Australia15 but lower than that in a study from the United States.7 The differences between this study and the latter study may be related to various factors influencing prevalence, as stated earlier.7,35,36 The higher PR of dementia in this study also might be related to lack of exclusion of prestroke dementia subjects. The mean APR to dementia in our stroke cohort agrees with the reported annual 3% incidence of dementia reported in the meta-analysis, indicating omparable development of new cases of psDem.22 The fluctuating prevalence of psDem over the assessment period (Table 2) is likely due to the changing frequency of attrition (Fig 1) and conversion to psDem. Stroke recurrence and conversion of psMCI to psDem also contributed to the rising prevalence of psDem in the

P value No CD (n 5 225)

psMCI (n 5 17)

psDem (n 5 39)

Age at first assessment, years, mean 6 SD Sex, n (%) Male Female Formal education, years, mean 6 SD Residence, n (%) Slum Nonslum Monthly family income, INR, n (%) ,2500 2501-5000 5000-10,000 .10,000 Family history of stroke, n (%) Yes No Hypertension, n (%) Yes No Diabetes, n (%) Yes No Smoking, n (%) Yes No Nonsmoking tobacco, n (%)* Yes No Alcohol, n (%) Yes No Age at FES, years, mean 6 SD Duration of disease, months, mean 6 SDy Number of strokes, n (%) Single Recurrent Side affected, n (%)z Left

62.1 6 13.15

68.8 6 9.34

74.5 6 8.30

,.001

.086

,.001

.256

108 (48.00) 117 (52.00) 5.4 6 4.83

11 (64.71) 6 (35.29) 5.8 6 4.11

27 (69.23) 12 (30.77) 5.4 6 5.28

.028

.021

.016

.763

.852

—

—

—

80 (35.56) 145 (64.44)

3 (17.65) 14 (82.35)

14 (35.90) 25 (64.10)

.320

—

—

—

30 (13.33) 126 (56.00) 66 (29.33) 3 (1.33)

1 (5.88) 14 (82.35) 2 (11.76) 0 (0.00)

5 (12.82) 16 (41.03) 18 (46.15) 0 (0.00)

.051

—

—

—

132 (58.67) 93 (41.33)

10 (58.82) 7 (41.18)

17 (43.59) 22 (56.41)

.211

—

—

—

155 (68.89) 70 (31.11)

12 (70.59) 5 (29.41)

25 (64.10) 14 (35.90)

.821

—

—

—

39 (17.33) 186 (82.67)

6 (35.29) 11 (64.71)

4 (10.26) 35 (89.74)

.076

—

—

—

68 (30.22) 157 (69.78)

4 (23.53) 13 (76.47)

14 (35.90) 25 (64.10)

.628

—

—

—

40 (17.78) 185 (82.22)

3 (17.65) 14 (82.35)

8 (20.51) 31 (79.49)

.918

—

—

—

28 (12.44) 197 (87.56) 57.2 6 12.70 58.5 6 54.57

2 (11.76) 15 (88.24) 64.3 6 9.20 45.3 6 50.51

3 (7.69) 36 (92.31) 69.1 6 7.51 66.3 6 57.34

.696

—

—

—

169 (75.11) 56 (24.89)

11 (64.71) 6 (35.29)

67 (29.78)

4 (23.53)

Overall

No CD vs psMCI

No CD vs psDem

psMCI vs psDem

,.001 .211

.051 —

,.001 —

.354 —

20 (51.28) 19 (48.72)

.008

.388

.004

.395

6 (15.38)

.185

—

—

— (Continued )

1239

Potential risk factor

COGNITIVE DYSFUNCTION AMONG STROKE SURVIVORS IN INDIA

Table 4. Univariate analysis of potential risk factors for psMCI or psDem considering the situation at baseline

S. DAS ET AL.

1240

Survival probability (%)

.014 ,.001 — ,.001 ,.001 —

.320 ,.001

— —

No CD vs psDem

80 60 40 20 0

,.001 ,.001 ,.001 ,.001 ,.001 ,.001

Abbreviations: INR, Indian rupees; SD, standard deviation. *Includes chewing tobacco and snuff. yTime from FES to assessment. zAccording to computed tomography (CT) reports; for the others, CT was not available, was normal, or showed involvement of both categories. xEleven SS with psDem at baseline could be evaluated with the BMSE, whereas none could be evaluated with the GDS.

.141 ,.001

5 (29.41) 8 (47.06) 215 6 3.32 39 6 3.08 225 6 6.63 32 (14.22) 139 (61.78) 27.2 6 2.43 1.4 6 2.15 14.5 6 8.36

15 (38.46) 19 (28.21) 186 6 4.39 85 6 1.68 —

— .240 6 (35.29) 6 (35.29) 98 (43.56) 68 (30.22)

19 (48.72) 5 (12.82)

6 (35.29)

Right Stroke type, n (%)z Infarct Hemorrhage Cortical atrophy, n (%)z Yes No BMSE score, mean 6 SDx EASI score, mean 6 SD GDS score, mean 6 SDx

65 (28.89)

8 (20.51)

No CD vs psMCI

0

Overall psDem (n 5 39) psMCI (n 5 17) No CD (n 5 225) Potential risk factor

Table 4. (Continued )

P value

psMCI vs psDem

100

100

200 300 Time (month)

400

500

Figure 2. Kaplan-Meier survival plots of survival trends in SSs with psDem (solid line), SS with psMCI (stippled line), and SS without CD (dotted line). The time axis denotes the time since FES. Censored observations are indicated by tick marks. Note that median survival in the latter 2 groups exceeded the duration of the study.

subsequent assessments. Approximately half of the SS with psDem developed dementia after recurrent stroke, comparable to rates reported in previous studies.33,37 In SS without recurrent stroke, other factors, such as underlying neurodegeneration, might have contributed to the evolution of dementia over the long term.33 MCI of vascular origin is considered a form of vascular cognitive impairment related to either cerebrovascular disease or its risk factors.38 Only SS with symptomatic stroke were included in this study, and here psMCI referred to CD not rising to the definition of dementia.38 Community-based studies on psMCI or its correlates after stroke are limited.5,15,39 The lower prevalence of psMCI in our study may be related to inclusion of the milder stroke cases in the community, in contrast to more severely affected hospital-based subjects. The importance of age as a risk factor for the conversion of psMCI to psDem was also reported in a previous study.15 The prevalence of psMCI among SS in this study was lower than that of MCI in a nondemented population from the same city.32 The overall rate of CD (including both psDem and psMCI) in our series at final assessment was 20.88% (Fig 1), lower than that reported in previous studies from developed countries (Table 1).5,9 This difference may be related to differences in defining criteria, available cohort size, age, and ethnicity among studies. The frequency of progression of psMCI to psDem in the present study is comparable to that in a previous study from Canada that documented an annual rate of conversion of MCI to dementia of 10%.39 A similar study from Singapore reported an annual conversion rate of 11%,9 comparable to that in the present study. Stroke recurrence and cortical atrophy indicate loss of cerebral mass and were identified on univariate analysis as risk factors for psDem in this study, in agreement with

COGNITIVE DYSFUNCTION AMONG STROKE SURVIVORS IN INDIA 7,11,33,40

previous studies. Table 3 indicates the influence of recurrent stroke in the development of CD. A higher frequency of stroke episodes is associated with an increased risk for dementia. psDem is considered a surrogate marker for underlying severe vascular disease leading to frequent stroke recurrence.33 Interestingly, socioeconomic differences was not associated with the risk of dementia in this study, in contrast to a previous study35; More over, an association of socioeconomic factors with stroke was not evident in a previous study from the same location.16 The SS with dementia in our series had less favorable survival, with a more than 2-fold greater risk of mortality than the SS without dementia. However, age might be an important confounder in inferring any association between dementia and death, given that mean age was .10 years older in SS with psDem than in SS without dementia. Previous studies have documented a 2- to 5-fold greater mortality in SS with psDem compared with SS without dementia, which could be independent of age and comorbidity.19,33 Survival was similar in the SS with psMCI and the SS without CD; median survival could not be calculated in either group, because .50% had survived by the end of the study. This study has several limitations. Difficulty in performing cognitive assessment of aphasic patients and the high attrition rate (nearly 44%) might have led to possible selection bias as well as underestimation of prevalence in the community. The high mortality rate in original stroke cohort meant that more severe cases who would have been more likely to suffer from dementia had they survived could not be assessed. This might have lowered the prevalence. Substantial sample loss due to various causes of 42%-46.5% has been reported in previous studies in both developed and developing countries.3,5 Our estimated prevalence of psDem was higher after correcting for attrition by the LOCF method, suggesting a selection bias. Conversely, our estimate of poststroke CD might was not related to have been lower had complete information about prestroke cognitive status been available. Although we sought such information qualitatively during the interviews, we did not use an objective scale, such as the Informant Questionnaire on Cognitive Decline in the Elderly. Our use of the BMSE for screening might have led to an underestimation of psMCI prevalence because this tool lacks detail on nonmemory items. Notwithstanding these limitations, this study from a developing country shows a higher prevalence of poststroke dementia than has been reported in many developed countries. In our series, the annual rate of progression of SS to psDem and the annual conversion of psMCI to dementia were comparable, suggesting that disease prognostic factors are not materially different. Improved stroke prevention can reduce stroke incidence and recurrence and help decrease the prevalence

1241

of poststroke CD. There is an urgent need for policy makers to plan for the institution of stroke preventive measures along with necessary cognitive rehabilitation strategies in India and other developing countries. Acknowledgment: We are deeply indebted to Prof. Mary Ganguli, Department of Psychiatry, Neurology and Epidemiology, University of Pittsburgh, for reviewing the manuscript and offering valuable suggestions. We thank the members of the Task Force on Neurological Disorders of the Indian Council of Medical Research for their guidance, and Gita Rudra for data entry and secretarial support.

References 1. Das S, Hazra A, Ray BK, et al. Burden among stroke caregivers: Results of a community-based study from Kolkata, India. Stroke 2010;41:2965-2968. 2. Tatemichi TK, Paik M, Bagiella E, et al. Dementia after stroke is a predictor of long-time survival. Stroke 1994; 25:1915-1919. 3. Tang WK, Chan SS, Chiu HF, et al. Frequency and determinants of poststroke dementia in Chinese. Stroke 2004; 35:930-935. 4. Madureira S, Guerreiro M, Ferro JM. Dementia and cognitive impairment three months after stroke. Eur J Neurol 2001;8:621-627. 5. Patel MD, Coshall C, Rudd AG, et al. Cognitive impairment after stroke: Clinical determinants and its associations with long-term stroke outcomes. J Am Geriatr Soc 2002;50:700-706. 6. Linden T, Skoog I, Fagerberg B, et al. Cognitive impairment and dementia 20 months after stroke. Neuroepidemiology 2004;23:45-52. 7. Ivan CS, Seshadri S, Beiser A, et al. Dementia after stroke: The Framingham study. Stroke 2004;35:1264-1268. 8. Rasquin SM, Lodder J, Ponds RW, et al. Cognitive functioning after stroke: A one-year follow-up study. Dement Geriatr Cogn Disord 2004;18:138-144. 9. Tham W, Auchus AP, Thong M, et al. Progression of cognitive impairment after stroke: One-year results from a longitudinal study of Singaporean stroke patients. J Neurol Sci 2002;203-204:49-52. 10. del Ser T, Barba R, Morin MM, et al. Evolution of cognitive impairment after stroke and risk factors for delayed progression. Stroke 2005;36:2670-2675. 11. Lowery K, Ballard C, Rodgers H, et al. Cognitive decline in a prospectively studied group of stroke survivors, with a particular emphasis on the .75s. Age Ageing 2002; 31(Suppl 3):24-27. 12. Aharon-Peretz J, Daskovski E, Mashiach T, et al. Natural history of dementia associated with lacunar infarctions. J Neurol Sci 2002;203-204:53-55. 13. Sundar U, Adwani S. Post-stroke cognitive impairment at 3 months. Ann Indian Acad Neurol 2010;13:42-46. 14. Prencipe M, Ferretti C, Casini AR, et al. Stroke, disability and dementia: Results of a population survey. Stroke 1997;28:531-536. 15. Srikanth VK, Anderson JF, Donnan GA, et al. Progressive dementia after first-ever stroke: A community-based follow-up study. Neurology 2004;63:785-792. 16. Das SK, Banerjee TK, Biswas A, et al. A prospective community-based study of stroke in Kolkata, India. Stroke 2007;38:906-910.

1242 17. Almeida OP, Xiao J. Mortality associated with incident mental health disorders after stroke. Aust NZ J Psychiatry 2007;41:274-281. 18. Hobson P, Meara J. Cognitive function and mortality in a community-based elderly cohort of first-ever-stroke survivors and control subjects. J Stroke Cerebrovasc Dis 2010;19:382-387. 19. Barba R, Morin MD, Cemill an C, et al. Previous and incident dementia as risk factors for mortality in stroke patients. Stroke 2002;33:1993-1998. 20. Desmond DW, Moroney JT, Sano M, et al. Mortality in patients with dementia after ischemic stroke. Neurology 2002;59:537-543. 21. Brodaty H, Altendorf A, Withall A, et al. Mortality and institutionalization in early survivors of stroke: The effects of cognition, vascular mild cognitive impairment, and vascular dementia. J Stroke Cerebrovasc Dis 2010;19:485-493. 22. Pendlebury ST, Rothwell PM. Prevalence, incidence, and factors associated with pre-stroke and post-stroke dementia: A systemic review and meta-analysis. Lancet Neurol 2009;8:1006-1018. 23. Das SK, Biswas A, Roy T, et al. A random sample survey for prevalence of major neurological disorders in Kolkata. Indian J Med Res 2006;124:163-172. 24. Sen V. Census of India 2001. Series 20. West-Bengal provincial population totals (vol 2). New Delhi: Government of India, 2001:P-XX. 25. Ganguli M, Ratcliff G, Chandra V, et al. A Hindi version of MMSE: The development of a cognitive screening instrument for a largely illiterate rural elderly population in India. Int J Geriatr Psychiatry 1995;10:367-377. 26. Das SK, Banerjee TK, Mukherjee CS, et al. An urban community-based study of cognitive function among non-demented elderly population in India. Neurol Asia 2006;11:37-48. 27. Folstein MF, Folstein SE, McHugh PR. Mini Mental State: A practical method for grading the cognitive state of patients for the clinicians. J Psychiatr Res 1975;12:189-198.

S. DAS ET AL. 28. Morris JC. The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology 1993;43:2412-2414. 29. Vas CJ, Pinto C, Panikker D, et al. Prevalence of dementia in an urban Indian population. Int Psychogeriatr 2001; 13:439-450. 30. Hatano S. Experience from a multicentre stroke register: A preliminary report. Bull World Health Organ 1976; 54:541-553. 31. American Psychiatric Association. Diagnostic and statistical manual of mental disorders, Third ed rev. Washington, DC: American Psychiatric Association, 1987. 32. Das SK, Bose P, Biswas A, et al. An epidemiologic study of mild cognitive impairment in Kolkata, India. Neurology 2007;68:2019-2026. 33. Leys D, Henon H, Mackowiak-Cordoliani MA, et al. Poststroke dementia. Lancet Neurol 2005;4:752-759. 34. Streiner DL. The case of the missing data: Methods of dealing with dropouts and other research vagaries. Can J Psychiatry 2002;47:68-75. 35. Pohjasvaara T, Erkinjuntti T, Vataja R, et al. Dementia three months after stroke: Baseline frequency and effect of different definitions of dementia in the Helsinki Stroke aging memory study (SAM) cohort. Stroke 1997;28:785-792. 36. Desmond DW, Moroney JT, Paik MC, et al. Frequency and clinical determinants of dementia after ischemic stroke. Neurology 2000;54:1124-1131. 37. Pandlebury ST, Rothwell PM. Risk of recurrent stroke, other vascular events and dementia after transient ischemic attack and stroke. Cerebrovasc Dis 2009;27(Suppl 3):1-11. 38. Hachinski VC, Bowler JV. Vascular dementia. Neurology 1993;43:2159-2160. 39. Wentzel C, Rockwood K, MacKnight C, et al. Progression of impairment in patients with vascular cognitive impairment without dementia. Neurology 2001;57:714-716. 40. Tatemichi TK, Foulkes MA, Mohr JR, et al. Dementia in stroke survivors in the stroke data bank cohort: Prevalence, incidence, risk factors and computed tomography. Stroke 1990;21:858-866.