Disability after stroke : a systematic review

Introduction: Stroke is the most common cause of disability in Western countries, yet there is no consensus in the literature on how to measure and describe disability from stroke. Objective: To conduct a systematic literature review on disability in stroke survivors. Method: Observational studies published in the PubMed, LILACS and SciELO online databases were selected, to evaluate disability in adults and in the elderly after stroke in the period 2002–2012. The Downs and Black checklist for non-randomized studies was used to assess the quality of the articles. Results: 212 articles were found from which 16 were selected to compose the study. The mean age of participants was 67 years, and disability affected 24% to 49% of the population evaluated. With regard to measurement instruments, 31% of the studies analyzed presented results of disability by means of the modified Rankin Scale; 19% by means of the World Health Organization’s International Classification of Functioning, Disability and Health; 19% by means of Katz’ Index of Independence in Activities of Daily Living; 12.5% by means * JFC: MSc, e-mail: jufabres@ig.com.br RLM: PhD, e-mail: renato.morelato@emescam.br HPP: MSc, e-mail: pereira_hudson@hotmail.com ERAO: PhD, e-mail: elizabete_regina@hotmail.com


Introduction
Stroke is the term used for a group of diseases that have abrupt onset and cause neurological damage, and is the most common cause of disability in Western countries (1,2,3).It is estimated that 25% to 74% of the 50 million stroke survivors in the world have some physical, cognitive or emotional deficiency, and require partial or complete assistance to carry out activities of daily living (ADL) (4).In recent years, advances in treatment of stroke reduced mortality rates, which, added to the phenomenon of aging, resulted in growth of the population that survives stroke (4)(5)(6)(7).
Some theoretical models were created in order to explain and establish relationships between the concepts of deficiency, disability and need for assistance (8,9,10).In one study that reviewed some of these models, Alves et al. (10) defined disability as a dynamic process that encompasses physical, mental and emotional conditions.Functional disability is understood as one part of this process, and is considered online course recommendation of a prior search for systematic reviews on the topic in order to evaluate the need for the study.The search for reviews was performed in the Cochrane library and online using MEDLINE, with the following keywords: "systematic review", "stroke" and "disabled person" (14).

Search strategy
The search strategy for this review included research in three online databases: MEDLINE via PubMed (National Library of Medicine and National Institutes of Health); LILACS (Latin American and Caribbean Health Sciences Literature) and SciELO (Scientific Electronic Library Online).The key words used were matched in English and Portuguese, and registered in the National Library of Medicine's controlled vocabulary thesaurus (MeSH) and the Health Sciences Descriptors (Decs) trilingual thesaurus: "stroke" and "disabled persons" in English, and "acidente vascular cerebral" and "pessoa com deficiência" in Portuguese.The terms "disability" and "functional disability" were not used as descriptors because they were not registered in the Decs.The search period was January 2002 through December 2012, with no restriction on language.

Selection of studies
Titles and abstracts were read for initial selection of the articles identified.The full articles were then read, and those that met the inclusion criteria were included in the data collection phase, performed by means of a standardized form (Annex 1).Identification and selection of the studies was performed by two researchers working independently.

Eligibility criteria
The articles selected were observational and cross-sectional or cohort, with the aim of evaluating the disability or level of assistance required in adults and elderly individuals after suffering stroke.Studies that evaluated other outcomes in addition to disability were included; however, those that associated stroke with other clinical situations, or that compared specific interventions and methods of rehabilitation, were excluded.Articles that were not published in English, Portuguese or Spanish, and those that were not found even after attempting to contact the author were also excluded.

Quality assessment
The studies included were initially divided according to study design, and the longitudinal studies were evaluated for quality by adapting the Downs and Black checklist for non-randomized studies, using only the 13 items relating to assessment of cohort studies.Adaptation of the Downs and Black instrument has already been used in other reviews in order to identify relevant methodological features of observational studies (15,16,17).

Results
The initial search yielded 212 documents; of these, 40 were selected.After reading each article in full, 24 were excluded for not meeting the eligibility criteria, leaving a total of 16 articles analyzed and submitted to the data collection phase of this review (Figure 1).
Among the studies analyzed, 56.25% were published since 2008.The articles were listed in descending order by year of publication in Table 1, in which general data such as author/year, objective, study design, sample, participant age, instruments and results were summarized.Eight studies were cross-sectional, two of which presented results from the same Italian population.Eight studies were prospective longitudinal, of which two presented the results from one Australian cohort five and ten years after stroke.
Sample sizes ranged from 13 to 1233 subjects, with the mean being 254 participants of both sexes.The studies included adult populations of various age groups, and the mean participant age was 67 years.The moment of evaluation ranged from three months to ten years after stroke.Diagnostic criteria for definition of cases had some differences between studies: 68.7% conducted clinical examination and the patients were classified according to the International Classification of Diseases (ICD), and 50% had the diagnosis reviewed by image examination.Three studies applied scales for disease classification: Lo et al. (18)  Three studies did not mention diagnostic criteria used (6,17,20).Four articles (25%) set out to evaluate quality of life (QOL) in addition to disability, and found strong positive association between the physical domain of QOL instruments and disability (5,7,17,19).Of the studies analyzed, 31% (five articles) presented the result of disability by means of the modified Rankin Scale (mRS) (17,20,21,22,23); 19% (three articles) by means of the International Classification of Functioning, Disease and Health (ICF), one of which used the World Health Organization Disability Assessment Schedule (WHODAS 2.0) (7,24,25); 19% (three articles) by means of the Katz Index of Independence in Activities of Daily Living (6,26,27); 12.5% (two articles) by means of the London Handicap Scale (18, 28); 12.5% (two articles) by means of the Barthel Index (5,19); and 6.25% (one article) according to the Functional Independence Measure (FIM) (29).
Considering functional disability as an indicator of disability measured by means of instruments such as the Katz and Barthel scales and the mRS, the result of 12 of the 16 selected articles can be assembled as follows: in the first two years after stroke, disability affected 24% to 49% of the population evaluated; three to five years after stroke, that number was between 26% and 47%; and ten years after, 46% of the population had some level of disability.
The evaluation of quality applied to the cohort studies had a mean methodological score of 10.5, and the studies that obtained the highest score were Hardie et al. (23) and Lin et al. (29).The score of each study can be found in Table 1.The studies applied various measurement instruments.Appelros et al. (20) used the Barthel Index, but presented the result of the disability by means of the mRS, concluding that 37% of the cohort was disable (mRS 3-6).Martins et al. (17) used the same methodology, and concluded that 36% of the sample was independent in Basic Activities of Daily Living (BADL) through Barthel Index, but 47% had disability (mRS ≥ 3).
Hong et al. ( 22) applied the disability-adjusted life year (DALY) measure to survivors of stroke, and used the results of the mRS for the calculation.As 35% of participants were disabled (mRS 3-6), the mean number of years of life lost due to disability was 3.82 (CI 95% 3.68 to 3.96).Hankey et al. ( 21) and Hardie et al. ( 23) also used the modified Rankin Scale (mRS) to present the results of disability.
The other studies used the BI to classify level of dependence, including Lo et al. (18), who showed results of disability with the London Handicap Scale, but classified dependence in two levels: 24% had BI < 14, which is dependency in more than one BADL, and 30% were in the Barthel range 15-19.The Australian studies by Gall et al. ( 28) and Hankey et al. (21) evaluated people five years after the first stroke, and also used other instruments to present disability; however, they applied the BI and defined as dependent those subjects who had a score < 20, and found 36% and 47% of their subjects dependent in some BADL, respectively.26) used the Katz Index in their reviews, and respectively found that 45%, 41% and 46% of their samples was dependent on at least one of the six activities evaluated.Lin et al. (29) used the Functional Independence Measure (FIM), and classified 59% of the sample with severe and very severe disability (mean score of 65), and 41% with moderate and mild disability (mean score of 87).In addition to evaluating the need for assistance with BADL and mobility, the FIM also evaluates communication and cognition.
The International Classification of Functioning, Disability and Health (ICF) was used in three studies (7,24,25).Quintas et al. (24) showed that body function problems most frequently reported by subjects that suffer from stroke are: difficulties with memory (more than 50% of the sample), attention (approximately 50%) and muscle function strength (50% of the sample).Furthermore, 47% reported moderate to severe gait impairment.Martins et al. (25) observed that 34% of subjects had difficulty in functions related to movement, and 52% had limitations in activities and participation, especially in domestic activities.
This result corroborates Alves (10), who also showed absence of a well-defined standard to measure disability.The terms "disability" and "level of dependency" were constantly taken alternatively in the articles, instruments such as Barthel and Katz, which were developed to assess level of dependence in ADL, were used to present the results of disability.Guccione (30) and Alves (31) point out that these scales of difficulty and dependence are good instruments for measuring functional disability, which is an aspect of disability.This result shows that there is a preference in the literature for measuring disability according to physical performance in daily activities, a result already observed by Cerniauskaite et al. (7).However, the approach of the ICF warns that disability cannot only be evaluated by dependence in ADL.The WHODAS 2.0, a disability assessment scale created by the WHO, evaluates the individual's functioning in six areas of activity: understanding and communicating, physical mobility, self-care, living with people, life activities and participation in society.It seems to be an instrument capable of covering all domains of disability, but because of its recent publication, it had little reference in the studies (32).

Disability in time
Functional capacity decreases with time, whereas the risk of developing a chronic condition increases (33).The debilitating process proceeds through the evolution of the chronic condition associated with other factors such as lifestyle, behavior, and biological, social and demographic characteristics (34).
Four cohorts followed the patients during the first year after stroke.Petrea et al. (27) found that three months after the event, the level of incapacity was greater than when compared after six months.Lo et al. ( 18) had a similar result, and concluded that its participants were less disabled after one year than in the third month, suggesting that functionality can improve between the third and twelfth month after stroke.This finding is consistent with previous articles that report a recovery between the third and sixth month after stroke (35,36).
Most of the cross-sectional studies did not define the exact moment after stroke, and assessed subjects at various times (5,6,7,24), while 19%, four cohorts, investigated long-term disability.Hankey et al. (21) stated that disability and institutionalization are the most common long-term results, affecting one-third and one-seventh of the survivors respectively.These authors also consider that the most important modifiable prognostic factors for these results are the low levels of physical activity and recurrence of stroke.Patel et al. (19) found that five years after stroke, 36% of the survivors of their cohort were dependent in some way.Gall et al. (28) found a similar result, where 47% of the survivors were dependent for some ADL after five years.

Evaluation of methodological quality and limitations of the study
The methodological variation between the studies, such as period evaluated, measurement instruments used and the form of data presentation made it difficult to gather the results of the 16 articles.Although the eight longitudinal studies evaluated had good methodological quality within the criteria of Downs and Black, the studies with cross-sectional design were not evaluated due to lack of scale or criterion for this purpose, which can be considered a limitation

Variables associated with disability and types of analysis
The results of four cohorts brought variables that were predictive or associated with disability.Gall et al. (28), by means of multivariate regression analysis, found that five years after stroke, increasing age and the recurrence of stroke (P < 0.01) were associated with greater disability.The Perth cohort also identified advanced age (OR = 5.7 in the age range 75-84; CI 95% 1.7-18.9),stroke recurrence (OR = 14.4,CI 95% CI 3.3-63), and moderate hemiparesis (OR = 3.4; CI 95% 1.2-9.5)as prognostic factors of disability, using logistic regression analysis adjusting the sex (23).
Using the logistic regression model, Lin et al. (29) found that bilateral compromise (OR = 10.8, p < 0:02) and low FIM score (OR = 7.6, p < 0.008) are the main variables associated with disability.Lo et al. (18) used the multilevel analysis and concluded that depression, advancing age and level of dependence are strongly associated with disability and severity (p < 0.001).
Petrea et al. (27) explored gender difference in post-stroke disability in Framingham's cohort, by means of bivariate analysis and logistic regression for variable adjustments considered risk factors for stroke according to Framingham.The authors found that women are four times more likely to be dependent prior to stroke than men (OR = 4.3, p < 0.01), and six months after stroke the chances of dependency remain higher in women (OR = 3.7, p < 0.01).
Five cohorts assessed the institutionalization rate of stroke survivors.Appelros et al. (20) aimed to define changes in living scenario, and found that prior to stroke, 13% lived in special housing, and one year after stroke that number had risen to 20%.The Perth cohort showed that five years after stroke, 17.7% of its members were institutionalized, and ten years after, 15% of the surviving members were institutionalized (21,23).Lo et al. (18) found that 19% of its population was institutionalized three months after stroke.The study that explored gender differences in the Framingham cohort found that women are 3.5 times more likely (p < 0.01) to be institutionalized six months after stroke than men (rate of 35% for women and 10% for men) (27). of this study.In addition, four articles were excluded from the study, two because they were not found, and two due to language.

Conclusion
This systematic review revealed heterogeneity in the literature when the subject is disability among stroke survivors.Most studies evaluated aspects of disability, failing to include all domains involved in this phenomenon.Furthermore, there is still no consensus as to the instrument to be used.
Wagering on functional disability as a reliable indicator of disability, it is possible to conclude that 24% to 49% of the population that survives stroke has some level of disability, which can vary according to age, stroke recurrence, time of evaluation and instrument applied.Most of the studies present disability by means of instruments that measure physical performance in activities of daily living, affirming functional disability as the most used indicator to define disability in people after suffering stroke.The Barthel Index was the instrument most used in the articles; however, the modified Rankin Scale was preferred to present results of disability.

Table 1 -
Characteristics of studies selected population.Strong correlation (ƿ = 0.84 and 0.75) in the physical function domains of the SF-36 with the areas of mobility and domestic tasks, respectively, of the WHODAS (p < 0.01).transfer, and 37% for walking.For men these percentages were 13% and 18%, respectively (OR = 2.3, OR = 1.91).Quality score: 11.

Table 1 -
Characteristics of studies selected(To be continued)

Table 1 -
Characteristics of studies selected Note: SF-36 = Short Form Health Survey; WHODAS 2.0 = World Health Organization Disability Assessment Schedule 2.0; ICF = International Classifi cation of Functioning, Disability and Health; ICD = International Classifi cation of Diseases; NIHSS = National Institute of Health Stroke Scale; mRS = modifi ed-Rankin Scale; DALY = disability adjusted life year; SIS = Stroke Impact Scale; LHS = London Handicap Scale; IDA = Irritability, Depression and Anxiety; MMSE = Mini-Mental State Examination; GDS = Geriatric Depression Scale; FAI = Frenchay Activities Index; COOP/WONCA = Primary Care Cooperative Information Project/World Organization of National Colleges Academies; MPI-S = Multidimensional Pain Inventory; MotrInd = Motricity Index.

Table 1 -
Characteristics of studies selected