Annals of Indian Academy of Neurology
ORIGINAL ARTICLE
Year
: 2020  |  Volume : 23  |  Issue : 4  |  Page : 515--521

Anemia as a predictor of functional disability in the early stage of ischemic stroke in a South Asian Population


Thashi Chang1, Praveen Weeratunga1, Thamal Vithanage2, Piumi Wijewickrama2, Sithara Kularathne2, Sachie Fernando2, Carukshi Arambepola3,  
1 Department of Clinical Medicine, Faculty of Medicine, University of Colombo; Department of Clinical Medicine, University of Colombo, National Hospital of Sri Lanka, Colombo, Sri Lanka
2 Department of Clinical Medicine, University of Colombo, National Hospital of Sri Lanka, Colombo, Sri Lanka
3 Department of Clinical Medicine, University of Colombo, National Hospital of Sri Lanka; Department of Community Medicine, University of Colombo, Colombo, Sri Lanka

Correspondence Address:
Praveen Weeratunga
Department of Clinical Medicine, Faculty of Medicine - University of Colombo, 25, Kynsey Road, Colombo - 00800
Sri Lanka

Abstract

Background: Reduced hemoglobin concentration has an adverse impact on the ischemic penumbra in patients with ischemic stroke as it causes reduced oxygen delivery to neuronal tissue and predisposes to infarct expansion. There is a paucity of data on the impact of anemia on early functional outcomes. Aims: To determine the association of anemia on early functional outcomes in a cohort of patients with ischemic stroke. Methods: This prospective study was conducted among 190 participants with acute ischemic stroke presenting to the National Hospital of Sri Lanka. Data were collected on socio-demographic determinants, clinical presentation, co-morbidities, subtype of stroke, and stroke severity (NIHSS score). Early functional outcomes were assessed by the Modified-Rankin-Score (mRS) and Barthel index (BI) within 48 h of the onset. Anemia was defined as Hb <13 g/dl in males and <12 g/dl in females. Results: The mean age of the population was 62.4 years (SD = 11.8). Most participants (75.8%) were males. Anemia was noted in 56.4% of the total study population (59.0% males; 56.5% females) with a mean Hb of 11.7 g/dl. A total of 20% of patients had moderate to severe stroke severity as defined by an NIHSS of ≥16. Functional status assessment revealed that 67.9% had mRS <3 and 85.8% had BI <75. Furthermore, 85.8% had a composite MRS <3 and/or BI 75. Univariate analysis revealed that anemia was significantly associated with “moderate-severe” functional disability. On logistic regression analyses, this retained significance when the functional disability was assessed by mRS >3 (adjusted OR = 2.36; 95% CI = 1.1–5.1). Receiver operator characteristics (ROC) curves indicated a Hb% of 10.65 g/dl as the cut-off that would predict stroke-related disability assessed by mRS >3 [sensitivity = 86.7%; specificity = 34.2%; and AUC = 0.659 (P < 0.0001)]. Conclusions: Anemia is an independent determinant of poor functional disability in early acute ischemic stroke.



How to cite this article:
Chang T, Weeratunga P, Vithanage T, Wijewickrama P, Kularathne S, Fernando S, Arambepola C. Anemia as a predictor of functional disability in the early stage of ischemic stroke in a South Asian Population.Ann Indian Acad Neurol 2020;23:515-521


How to cite this URL:
Chang T, Weeratunga P, Vithanage T, Wijewickrama P, Kularathne S, Fernando S, Arambepola C. Anemia as a predictor of functional disability in the early stage of ischemic stroke in a South Asian Population. Ann Indian Acad Neurol [serial online] 2020 [cited 2020 Oct 25 ];23:515-521
Available from: https://www.annalsofian.org/text.asp?2020/23/4/515/267878


Full Text



 Introduction



In spite the recent surge of therapeutic options in treating acute stroke that has shifted from time-locked thrombolytic therapy to tissue-based mechanical thrombectomy that has extended the therapeutic window up to 24 h,[1] stroke remains a leading cause of adult disability.[2],[3] Among the many reasons as to why successful therapeutic reperfusion of acute ischemic strokes does not always result in complete recovery, anemia has been considered a possible factor. On one hand, it is shown that anemia is associated with increased stroke severity by interfering with the salvation of neurons in the ischemic penumbra by depleting the oxygen-carrying capacity of the blood, which could result in poor functional stroke outcomes, while on the other hand, anemia may have a direct effect on functional disability.[4] Recent evidence according to retrospective analyses clearly informs that anemia is a predictor of death following a stroke.[5],[6] However, subsequent prospective studies have found an uncertain relationship between anemia and stroke outcomes when adjusted for confounders such as smoking and blood pressure.[7] Furthermore, the main focus of previous studies has been on the impact of anemia on mortality related to stroke, whereas studies conducted on the impact of anemia on functional outcomes are scarce. Studying this relationship at an early time point is relevant because long-term functional outcomes following stroke are largely determined by the level of disability at the early stage of the disease. This has implications on stroke management particularly because the presence of anemia in the individual patient is likely to determine the individual's prognosis after stroke, whereas the prevalence of anemia is likely to impact on the outcome of acute reperfusion therapies in the population at large.

Sri Lanka is a low- and middle-income country with a population of approximately 21 million and is estimated to have one of the highest age-standardized stroke mortality and stroke mortality-to-incidence ratios.[8] Although the prevalence of stroke in urban Sri Lanka lies between high-income and low- and middle-income countries, it remains a major cause of long-term disability in Sri Lanka.[9] The prevalence of anemia among non-pregnant adult females in Sri Lanka is approximately 39%.[10] Although centers for stroke thrombolysis have been established in the free state health service in every province in Sri Lanka, the outcome of reperfusion therapy is likely to be dampened by the high prevalence of anemia in the population.

Given the relatively high prevalence of stroke and anemia in Sri Lanka, this study aimed to determine the impact of anemia on functional outcome of stroke in the early stage of disease in a high anemia prevalence setting while adjusting for confounding factors.

 Methods



The authors declare that all supporting data are available within the article. Ethics approval was obtained from the Ethics Review Committee, Faculty of Medicine, University of Colombo. July 2014.

Patients and setting

A hospital-based cross-sectional analytical study was conducted among patients aged 18 years or more with new-onset acute ischemic stroke who were admitted to the National Hospital of Sri Lanka. This is the apex tertiary referral hospital in the country located in the district of Colombo, which has the highest population density (3,300/km 2) with 2.3 million inhabitants. Patients are referred to this hospital from Colombo and all other districts in Sri Lanka. We identified the eligible patients for the study according to the World Health Organization (WHO) definition of stroke as “a focal (or at times global) neurological impairment of sudden onset, and lasting more than 24 hours, and of presumed vascular origin”,[11] whereas hemorrhage and alternative diagnoses for the patients' neurological deficits were excluded using cranial computerized tomography (CT). A normal cranial CT or an area of cerebral hypodensity corresponding to the neurological deficit was considered consistent with an acute ischemic stroke. Patients with hematological disorders other than anemia were excluded. The sample required was calculated using the equation for detecting a given odds ratio in a cross-sectional analytical study design,[12] considering an anticipated odds ratio of anemia (exposure) for functional disability in stroke (outcome) of 2.09 from the previous study;[13] probability of anemia prevalence (exposure) in the general population (without disease) of 0.34 from Demographic and Health Survey Sri Lanka 2007;[14] 95% level of confidence; and relative precision of 50%. Assuming that 35% of the stroke patients would be identified as having a functional disability, the final sample size was 191. The sample of patients was obtained consecutively from the patient admission registers.

Assessments

Data were collected by two trained pre-intern medical graduates. An interviewer-administered questionnaire was used to record socio-demographic data, co-morbidities, and risk factors (hypertension, dyslipidemia, diabetes mellitus, ischemic heart disease, congestive cardiac failure, atrial fibrillation, history of stroke and transient ischemic attack, smoking, and alcohol consumption) from medical records and self-report. In patients unable to participate in an interview, the spouse or first-degree relatives were interviewed on their behalf.

Patients were evaluated systematically within 48 h of the onset of stroke for stroke type, subtype, stroke severity, and functional disability. Stroke was classified according to the Oxfordshire Community Stroke Project.[15] Stroke severity was assessed using the National Institute of Health Stroke Scale (NIHSS). Functional disability was assessed using the modified Rankin Scale (mRS) and the Barthel index (BI). Hemoglobin (Hb) concentrations were estimated on admission using a Beckmann Coulter auto-analyzer in accordance with the manufacturer's instructions.

Definitions

Anemia was defined using the WHO criteria, as Hb concentration less than 13 g/dl in men and less than 12 g/dl in women.[16] “Moderate-severe” level of stroke severity was defined by an NIHSS score of 16 or more.[17] “Moderate-severe” level of functional disability, the degree at which patients were considered to be functionally dependent, was defined from three independent classifications: mRS score more than 3; BI score less than 75; and composite index of mRS more than 3 and/or BI score less than 75. The mRS of 3 has been shown to be the corresponding score for a BI of 75 and recommended as the optimal cut-off values for defining stroke functional outcomes.[18]

Data analysis

Univariate analysis (analysis repeated for each classification on functional disability) was performed first to determine the factors associated with “moderate-severe” level of functional disability at the early stage of the disease, using Chi-square test with significance tested at 0.05 level. To adjust for confounders, logistic regression (LR) analysis (analysis repeated for each classification on functional disability) was performed, in which the dependent variable was the level of functional disability, whereas the independent variables were those significant in the univariate analysis. Backward LR method was utilized in model generation (exit at a significance level of 0.1). Adjusted odds ratio (OR) and 95% CI were derived for the variables in the final models.

Using the functional disability outcome that showed a significant association with anemia in the final LR model, optimal cut-off values of Hb that predict the functional disability in stroke were derived using receiver operator characteristic (ROC) curves. The area under the curve (AUC) determined the goodness of fit of the ROC curve.

Further, correlations of the Hb concentration with functional disability (according to mRS and BI scores) were assessed using Spearman correlation coefficient (rs). Further, simple regression analysis was carried out to assess the performance of Hb concentration (g/dl - X-axis) as a predictor of the functional disability (BI or mRS scores - Y-axis) of patients (dose-response effect). The analysis generated model summary statistics, parameter estimates, and a regression equation to describe the statistical relationship between the predictor variable and response variable.

Ethical considerations

Ethics approval was obtained from the Ethics Review Committee of the Faculty of Medicine, University of Colombo, Sri Lanka. Written informed consent was obtained from the patient or in incapacitated patients, from the next-of-kin/guardian.

 Results



Patient characteristics

The study recruited 190 participants with new-onset ischemic stroke. The majority were males (n = 144, 75.8%). Mean age of the sample was 62.4 years (SD = 11.8), ranging from 24 to 85 years. Of the stroke subtypes, total anterior circulation infarction (TACI) and lacunar circulation infarction were the most common. The most common traditional risk factors for stroke were hypertension (62.6%), smoking (52.1%), diabetes mellitus (47.4%), and dyslipidemia (36.8%). Of all patients, 20% indicated “moderate-severe” level of stroke severity. Mean Hb was 11.7 g/dl (SD = 2.2) [men = 11.7 (SD = 2.3); women = 11.65 (SD = 1.9)].A total of 111/190 (58.4%) patients had anemia (men = 59.0%; women = 56.5%). The basic characteristics of patients by their anemic status are shown in [Table 1]. None of the characteristics were significantly different between the anemic and non-anemic patients (P > 0.05) except past stroke (P = 0.001) and family history of stroke (P = 0.01).{Table 1}

Functional outcome in early stage of disease among patients with stroke

As shown in [Table 2], “moderate-severe” level of functional disability was found in 67.9% according to the mRS classification, and 85.8% according to the BI classification. When both classifications were considered together (mRS >3 and/or BI <75), it was 86.3%.{Table 2}

Association of anemia with functional disability of stroke

In the univariate analyses performed for all three classifications of functional disability [Table 3], a significantly higher proportion of “moderate-severe” level of functional disability was seen among anemic patients than that of the non-anemic ones [Table 3]. Other variables significantly associated with such disability were history of stroke, TACI stroke subtype, and “moderate-severe” level of stroke severity.{Table 3}

In the logistic regression analyses performed to identify the role of anemia on functional disability independent of other factors [Table 4], anemic state, history of stroke, stroke subtype, and stroke severity were included as independent variables. Anemia remained a significant factor associated with the functional disability when assessed according to mRS >3 (adjusted OR = 2.36; 95% CI = 1.1–5.1), in contrast to the marginal significance observed when assessed according to composite index (adjusted OR = 2.19; 95% CI = 1.01–4.75) and no relationship according to BI (P > 0.05).{Table 4}

ROC curves were generated to identify the Hb concentration, which would predict “moderate-severe” functional disability based on mRS >3. The optimal cut-off Hb concentration was computed as 10.65 g/dl, which had a sensitivity of 86.7% and a specificity of 34.2% [Figure 1]. The AUC was 0.659 (P < 0.0001).{Figure 1}

Prediction of functional disability of stroke using hemoglobin concentration

The average mRS and BI scores on functional disability were 3.63 (SD = 1.19) and 38.3 (SD = 30.6), respectively. These scores showed significant correlations with the Hbconcentration (mRS: rs= –0.2; P = 0.003; and BI: rs = 0.27; P = 0.000). As shown by significant beta coefficients in the regression models [Table 5], Hb concentration was a significant predictor variable of functional disability. Both models were statistically significant (mRS model: F = 8.01; P = 0.005; BI model: F = 14.72; P = 0.000). The derived regression equations are as follows:{Table 5}

Functional disability based on mRS score = 4.895+ –0.108* Hb concentration

Functional disability according to BI score = –4.962 + 3.688* Hb concentration

 Discussion



We found that almost two-third of our study population presenting with acute ischemic strokes were anemic. Furthermore, anemia was an independent determinant of poor functional outcome assessed using the mRS or composite score classifications. Hb concentration less than 10.65 g/dl was associated with worse stroke-related functional outcome, whereas a significant dose-response relationship was noted between Hb concentration and functional disability scores. Our results also showed that a history of stroke and the stroke subtype of total anterior circulation infarct were independent determinants of poor functional outcome. Although stroke severity per se was an independent risk factor of functional outcome, only a fifth of our study population had moderate to severe strokes. None of our patients had received thrombolytic therapy as this study was done prior to provision of thrombolysis in the free National Health Service in Sri Lanka.

A meta-analysis of 13 cohort studies has shown anemia to be an independent risk factor of increased mortality following both ischemic and hemorrhagic strokes.[6] However, studies on the impact of anemia on functional outcome in early stroke are scarce and are often confounded by other risk factors such as hypertension, diabetes mellitus, and dyslipidemia or prognostic determinants such as stroke severity, previous strokes, and stroke subtypes. Our study provides evidence for an independent and significant effect of anemia on the functional outcome of acute ischemic stroke in its early stage, which is likely to determine long-term outcome. Our results were significant when the functional status was measured either by mRS or compositely with mRS with or without BI. However, the unfavorable relationship of functional outcome and anemia was not significant when the functional status was measured by BI alone. This may have been because mRS is a measure of global disability, whereas BI is a summed measure of individual activities of daily living. Thus, our results also suggest that mRS is a more reliable measure of disability in the acute phase after stroke while assessing activities of daily living is more reliable in later stroke to evaluate rehabilitation outcomes.

There is a paucity of data on the functional outcomes of ischemic stroke in anemic patients in South Asia. This is concerning, especially owingto rising incidence of stroke in the region, as well as the heightened prevalence of anemia. Joshi et al. note a 42% prevalence of moderate anemia in a cohort of 50 patients with ischemic stroke from Nagpur, India.[19] Sukdev et al. describe an Indian patient with severe anemia presenting with ischemic stroke and explore the hypothesis of anemia as a risk factor in the pathogenesis.[20]

The relationship between the Hb concentration in blood and stroke outcome is complex. On one hand, increased hematocrit is likely to increase blood viscosity and interrupt normal flow dynamics, while on the other hand, reduced Hb is likely to impair tissue oxygenation. Lower Hb levels have been reported to be associated with increased stroke volumes as well as poorer stroke outcomes.[4],[21],[22] It has been suggested that in patients with acute ischemic stroke, anemia may cause aggravation of hypoxia in the penumbral lesions.[23] Consistent with this, recent retrospective analysis has found that moderate to severe anemia was associated with a poor outcome in spite successful revascularization with mechanical thrombectomy.[4] On the contrary, an inverse relationship between a higher hematocrit and cerebral blood flow has been reported.[24] Further complicating this relationship is the suggestion that anemia is a hyperkinetic state that alters endothelial function and predispose to thrombus formation and embolism. Furthermore, anemia induced erythropoietin secretion can lead to reactive thrombocytosis.[25] Thus, a U-shaped distribution has been proposed to denote the relationship between anemia and stroke outcome.[26] However, the optimal level of Hb that prevents infarction of critically ischemic cerebral tissue remains undetermined. Although we found a worse functional outcome below a cut-off Hb level of 10.65 g/dl, larger studies are required to determine a more accurate value with greater sensitivity and specificity.

The NIHSS score is considered a reliable measure of stroke severity and identified as a surrogate marker of lesion volume.[27] As expected, a greater stroke severity was associated with a worse functional outcome, but the occurrence of more severe strokes was not different between anemic and non-anemic patients in our study. It has been shown that the decrease in Hb after admission may be more relevant to stroke severity than baseline levels [28] suggesting that the development of collateral circulation may compensate in patients with anemia. Thus, our data support the observation that anemia acts independently to an initial lesion volume in determining the functional outcome of stroke.

The key finding of our study is the recognition of anemia as an independent risk factor of poor functional outcome following acute ischemic stroke. Our data also supports previous observations of a high prevalence of anemia among patients presenting with acute ischemic stroke. We focused on the acute phase of stroke presentation because this is the period in which therapeutic interventions have the greatest impact on prognosis. However, whether correction of anemia in the acute phase following stroke improves prognosis remains controversial.[25] However, our study lends justification for a randomized trial to test the therapeutic value of blood transfusions in patients with anemia presenting with acute ischemic stroke.

This study had a few limitations. The duration of anemia preceding stroke was not known. The causes or types of anemia was not determined, but sickle cell disease is exceptionally rare in Sri Lanka and is unlikely to have contributed significantly. Furthermore, the small sample size and cross-sectional design recruiting patients from a single center may have introduced statistical error and biases. A case-control design is not appropriate here, as functional disability was assessed using three independent classifications, thus leading to misclassification of cases (stroke patients with a disability).

 Conclusions



Our study highlights the role of anemia as an independent determinant of poor functional outcome in the early stage of acute ischemic stroke. Furthermore, our data reaffirms that anemia is common among South Asians presenting with stroke. Until evidence becomes available to support treatment of anemia in the acute phase of stroke, intervention should focus on screening and treating anemia in the general population to minimize the deleterious effects of it in the event of acute ischemic stroke.

Declaration of patient consent

The authors certify that they have obtained all appropriate patient consent forms. In the form, the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

References

1Nogueira RG, Jadhav AP, Haussen DC, Bonafe A, Budzik RF, Bhuva P, et al. Thrombectomy 6 to 24 hours after stroke with a mismatch between deficit and infarct. N Engl J Med 2018;378:11-21.
2Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med 2006;3:e442.
3Murray CJ, Vos T, Lozano R, Naghavi M, Flaxman AD, Michaud C, et al. Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990-2010: A systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012;380:2197-223.
4Akpinar CK, Gurkas E, Aytac E. Moderate to severe anemia is associated with poor functional outcome in acute stroke patients treated with mechanical thrombectomy. Interv Neurol 2018;7:12-8.
5Milionis H, Papavasileiou V, Eskandari A, D'Ambrogio-Remillard S, Ntaios G, Michel P. Anemia on admission predicts short- and long-term outcomes in patients with acute ischemic stroke. Int J Stroke 2015;10:224-30.
6Li Z, Zhou T, Li Y, Chen P, Chen L. Anemia increases the mortality risk in patients with stroke: A meta-analysis of cohort studies. Sci Rep 2016;6:26636.
7Kannel WB, Gordon T, Wolf PA, McNamara P. Hemoglobin and the risk of cerebral infarction: The Framingham Study. Stroke 1972;3:409-20.
8Feigin VL, Forouzanfar MH, Krishnamurthi R, Mensah GA, Connor M, Bennett DA, et al. Global and regional burden of stroke during 1990-2010: Findings from the Global Burden of Disease Study 2010. Lancet 2014;383:245-54.
9Chang T, Gajasinghe S, Arambepola C. Prevalence of stroke and its risk factors in Urban Sri Lanka: Population-based study. Stroke 2015;46:2965-8.
10Rajapaksa L, Arambepola C, Gunawardena N. Nutritional Status in Sri Lanka, Determinants and Interventions: ADesk Review 2006-2001. Colombo: UNICEF; 2011.
11World Health Organization. WHO STEPS Stroke Manual: The WHO STEPwiseApproach to Stroke Surveillance. Geneva: World Health Organization; 2006.
12Lwanga SK, Lemeshow S. Sample Size Determination in Health Studies: APractical Manual. Geneva: WHO; 1991.
13Tanne D, Molshatzki N, Merzeliak O, Tsabari R, Toashi M, Schwammenthal Y. Anemia status, hemoglobin concentration and outcome after acute stroke: A cohort study.BMC Neurol 2010;10:22.
14DCS. Demographic & Health Survey 2006-7. Colombo: Department of Census and Statistics; 2009.
15Bamford J, Sandercock P, Dennis M, Burn J, Warlow C. Classification and natural history of clinically identifiable subtypes of cerebral infarction. Lancet 1991;337:1521-6.
16WHO. Haemoglobin Concentrations for the Diagnosis of Anaemia and Assessment of Severity. Vitamin and Mineral Nutrition Information System. Geneva: World Health Organization; 2011 (WHO/NMH/NHD/MNM/11.1). Available from: http://www.who.int/vmnis/indicators/haemoglobin.pdf. [Last accessed on 2019 Jul 02].
17Adams HP Jr, Davis PH, Leira EC, Chang KC, Bendixen BH, Clarke WR, et al. Baseline NIH stroke scale score strongly predicts outcome after stroke: A report of the trial of org 10172 in acute stroke treatment (TOAST). Neurology 1999;53:126-31.
18Uyttenboogaart M, Stewart RE, Vroomen PC, De Keyser J, Luijckx GJ. Optimizing cutoff scores for the Barthel index and the modified Rankin scale for defining outcome in acute stroke trials. Stroke 2005;36:1984-7.
19Joshi A, Agrawal S. Iron deficiency anaemia as a risk factor for stroke. IOSR J Dent Med Sci (IOSR-JDMS) 2016;15:42-44.
20Sukdev M, Sweety K, Dhar M, Panda PK. Rare cause of stroke in young: Iron deficiency anemia and diabetic ketoacidosis. J Family Med Prim Care 2019;8:1775-7.
21Kimberly WT, Wu O, Arsava EM, Garg P, Ji R, Vangel M, et al. Lower hemoglobin correlates with larger stroke volumes in acute ischemic stroke. CerebrovascDisExtra 2011;1:44-53.
22Huang WY, Chen IC, Meng L, Weng WC, Peng TI. The influence of anaemia on clinical presentation and outcome of patients with first-ever atherosclerosis-related ischemic stroke. J Clin Neurosci 2009;16:645-9.
23Shahar A, Sadeh M. Severe anemia associated with transient neurological deficits. Stroke 1991;22:1201-2.
24Grotta J, Ackerman R, Correia J, Fallick G, Chang J. Whole blood viscosity parameters and cerebral blood flow. Stroke 1982;13:296-301.
25Kaiafa G, Savopoulos C, Kanellos I, Mylonas KS, Tsikalakis G, Tegos T, et al. Anemia and stroke: Where do we stand? Acta Neurol Scand 2017;135:596-602.
26Yassi N, Davis SM. Commentary: A cohort study of patients with anemia on admission and fatality after acute ischemic stroke. J Clin Neurosci 2013;20:43.
27Kasner SE. Clinical interpretation and use of stroke scales. Lancet Neurol 2006;5:603-12.
28Kellert L, Martin E, Sykora M, Bauer H, Gussmann P, Diedler J, et al. Cerebral oxygen transport failure?: Decreasing hemoglobin and hematocrit levels after ischemic stroke predict poor outcome and mortality: STroke: RelevAnt impact of hemoGlobin, hematocrit and transfusion (STRAIGHT)--an observational study. Stroke 2011;42:2832-7.