XVII. European Stroke Conference
Nice, France
XVII. European Stroke Conference
Nice, France
Poster Session: Meta-analysis and review
papers
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
01
Tirilazad for subarachnoid haemorrhage: a Cochrane revivew
S.H.Zhang
L.C.Wang
M.L.Liu
West China Hospital, Sichuan University
CHINA
Objectives To assess the safety and efficacy of tirilazad in aneurysmal subarachnoid haemorrhage. Methods A Cochrane systematic review of randomized controlled trials of Tirilazad for aneurysmal subarachnoid haemorrhage was performed. Trials of Tirilazad started within 10 days of subarachnoid haemorrhage onset, compared with placebo or open control were eligible for inclusion. Trials published before April 2007 were searched in the Cochrane Central Register of Controlled Trials, MEDLINE and EMBASE. Results Five trials involving 3797 patients were included. All the trials were double–blinded, vehicle-control trials and there were no significant heterogeneities among them. During treatment period, there were fewer symptomatic vasospasms occurred in tirilazad group than in control group (OR 0.80, 95%CI 0.69 to 0.93). However, there were no significant differences in death or overall poor outcomes (death, vegetative state or severe disability) between the two groups at the end of follow-up (OR 0.89, 95%CL0.74 to 1.06; OR1.02, 95%CL 0.88 to 1.19; respectively). Leukocytosis and prolongation of Q-T interval occurred significantly more frequently in treatment group were reported in only one trial (Tirilazad at dose of 15 mg/kg/day). There were no significant differences in infusion site diseases and other laboratory parameters between the two groups in all trials. Conclusions There is no evidence that tirilazad can reduce mortality or improve poor outcomes of the patients with subarachnoid haemorrhage. The routine use of tirilazad for aneurysmal subarachnoid haemorrhage can not be recommended on the base of current evidence.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
02
Triggers of ischaemic stroke: a systematic review and meta-analysis
V. Guiraud
M. Ben Amor
E. Touzé
J.L.Mas
Hôpital Sainte-Anne
FRANCE
Background Several studies have suggested that exposure to some acute factors, so-called “triggers”, could precipitate the onset of ischaemic stroke (IS), but the strenght of the relationships is uncertain. We performed a systematic review of the potential triggers of IS. Methods Pubmed [MEDLINE] and Embase searches were carried out by two independent reviewers to identify studies (cohort, case-control and case-crossover) published between January 1980 and September 2007. When possible, odds ratios (OR) were combined with a fixed effect. Methodological quality was assessed using a pre-defined criteria list. Results 17 case-control and 9 case-crossover studies identified 12 triggers. Significant associations were reported for acute alcohol abuse >40 g within the preceding 24 hours (4 studies, OR = 3.05; 95% CI 2.34−3.99; heterogeneity p=0.02), >150 g within the previous week (3 studies, OR = 2.69; 2.05−3.54; heterogeneity p=0.13) and infection within the previous week (5 studies, OR = 4.83; 3.30−7.07; heterogeneity p=0.2). Others studies reported significant associations between IS and negative emotions, anger, discontinuation of aspirin, general surgery, decrease in temperature and air pollution. Selection biases were noted in 47%. Only, 54% enrolled patients consecutively. Triggers were clearly defined in 96%; multivariate analyses were performed in 88% of all studies. Discussion Several acute risk factors for IS, so-called “triggers”, have been identified. Most studies have dealt with acute alcohol abuse and acute infection.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
03
Should stroke trials adjust their primary outcome for prognostic factors?
L.J.Gray
T. Collier
P.M.W.Bath
on behalf of the OAST Collaboration
Institute of Neuroscience, University of Nottingham, UK
UNITED KINGDOM
Background: Most stroke clinical trials have given neutral results. Sub-optimal statistical analyses may be failing to detect differences for effective interventions. Adjusting outcomes for baseline prognostic factors related to outcome may improve the efficiency of outcomes. Methods: Data from 30 stroke trials (28,316 patients) assessing functional outcome were included. The prognostic variables considered were age, sex and baseline severity. Unadjusted and adjusted ordinal logistic regression models were compared using simulated data from the 30 trials (5,000 simulations per trial). 3 levels of injected treatment effect was assessed, odds ratio (OR) 0.57, 0.74 and 0.95. The reduction in sample size gained from using the adjusted models was then calculated. Results: Across all 3 treatment effects a similar level of reduction in sample size was seen, median percentage reduction and inter-quartile range: OR=0.57: 35.1% (95% confidence interval (CI) 29.3 to 40.0), OR=0.74: 35.8% (95% CI 29.6 to 40.3) and OR=0.95: 32.5% (95% CI 30.0 to 39.5). As the treatment effect increased, the percentage of simulations where the treatment effect for the adjusted model was greater than for the unadjusted model also increased. Conclusion: Adjusting for key prognostic factors could reduce sample size by around 35%. Smaller trial sample sizes may help by reducing time to completion, trial complexity, and financial expense, and increase the number of trials that can be done.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
04
Retinal Abnormalities are Associated with Previous and Future Stroke but Data is Lacking on Stroke Subtypes: A Systematic Review
F.N.Doubal
P.E.Hokke
J.M.Wardlaw
University of Edinburgh
UNITED KINGDOM
Introduction: The exact aetiology of lacunar stroke is unknown. Retinal arterioles share many features with cerebral small vessels and could shed light on lacunar stroke aetiology. We systematically reviewed the literature to ascertain any association between retinal changes and stroke or its subtypes. Methods: We searched Medline and Embase for studies of retinal microvascular abnormalities (retinopathy, arteriolar and venular diameters, structural vascular changes, retinal vessel occlusion) and cerebrovascular disease (stroke, TIA, brain infarct on imaging) published between 1966 and October 2007. 2 reviewers extracted data; a 3rd checked and adjudicated disagreements. We pooled risk ratios using generic inverse variance methods to provide summary risk ratios (sRR). Results: We found 4461 titles of which 55 potentially relevant full text papers were assessed; 37/55 met the inclusion criteria representing 62,975 subjects, mean age 62 years;1,900 had a clinical cerebrovascular event (3%) and 993 (1.5%) had an asymptomatic infarct on imaging. 21/37 papers were from large population studies, 16/37 were hospital series. 21 papers reported retinal abnormalities and future cerebrovascular disease, 15 past cerebrovascular disease and 1 paper both future and past. Only 1 study compared lacunar and cortical ischaemic stroke (59 patients) and found no difference in retinal features. In pooled analysis, retinopathy was associated with future stroke risk (8 studies, (sRR 2.1, 95% CI 1.7-2.9) and previous stroke (7 studies, sRR 2.5, 1.4-4.3). Narrowed arterioles did not predict future stroke (3 studies, sRR 1.1, 0.5-2.3) but widened venules did (3 studies, sRR 2.2, 1.1-4.4). Retinal arteriolar embolism predicted future stroke (4 studies, sRR 2.0, 1.1-3.7). Conclusion: Retinal changes predict future and past stroke but there is virtually no data on stroke subtypes. The young age and few strokes limit the data. More studies are needed in large relevant stroke populations.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
05
Predictors of upper limb recovery following stroke: A systematic review
F. Coupar
P. Langhorne
P. Rowe
C. Weir
Academic Section of Geriatric Medicine, University of Glasgow
UNITED KINGDOM
Background: Upper limb hemiparesis is a common, persisting and disabling sequela of stroke. However, evidence for the effectiveness of interventions targeted at the upper limb remains inconclusive. Identification of reliable predictors of upper limb recovery would allow interventions to be better targeted at appropriate patients and thus potentially optimise upper limb rehabilitation. We carried out a systematic review of predictors of upper limb recovery. Methods: We completed searches in Medline, Embase, Amed, Cinahl and Cochrane CENTRAL databases. Articles were included if predictor variables were measured at baseline and related to an outcome of upper limb recovery at a future time point. Exclusion criteria included predictor variables relating to a particular treatment and outcome measurements of very specific upper limb impairments such as spasticity or pain. Results: To date two independent reviewers have identified 54 studies (over 6000 participants) that meet the inclusion criteria. Predictor variables which have been considered within these studies include; age, sex, lesion site, initial motor impairment, motor evoked potentials and somatosensory evoked potentials. Preliminary results indicate that the severity of the initial upper limb impairment is the most consistently reported and significant predictor of upper limb recovery. Discussion: Interpretation of these results is complicated by methodological factors including variations in study populations, upper limb motor outcome scales, timing of baseline and outcome assessments and predictors selected. The most important predictive factor for upper limb recovery following stroke appears to be the initial severity of motor impairment.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
07
Periodontal disease modifies incident stroke risk in adults without prior history of cardiovascular disease.
T. Gago
G. Alcoforado
J.M.Ferro
School of Dental Medicine, Universidade de Lisboa, Lisboa, Portugal
PORTUGAL
Background: Both stroke and periodontal diseases are significantly prevalent in the general population. Few studies have addressed the particular association between periodontal diseases/edentulism and stroke risk. The association relies on the concept of focal infection, producing a direct or indirect endothelial dysfunction. Methods: We performed a systematic review searching multiple database for cohort, case-control and cross-sectional studies. No time or language restrictions were used. Study selection and review was made by two reviewers. Quality assessment for heterogeneity analysis was performed. Generic invariance method with fixed effect models were used with RevMan 4.2.10 software. Results: Overall we identified six cohort studies; six case-controls and three cross-sectional studies. Periodontal disease was associated to incident stroke (RR=1.78; CI: 1.09-2.89; p=0.02), particularly to the ischemic type (RR=2.47; CI: 1.59-3.81; p<0.0001). Edentulousness was also associated to incident stroke (RR=1.35; CI: 1.07-1.69; p=0.01). Gingivitis was not statistically associated (RR=1.22; CI: 0.96-1.56; p=0.11) to incident stroke. Immunoglobulin production (IgA and IgG) against periodontopathogens was not statistically associated to incident stroke risk. Conclusion: Periodontal disease and edentulism are associated to stroke incidence. Periodontitis may be viewed as a risk marker for incident ischemic stroke in certain population sub-groups.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
08
Infarct volume in experimental animal models and clinical trials of acute stroke therapies
V.E.O'Collins
G.A.Donnan
M.R.Macleod
E.S.Sena
S.F.Cox
D.W.Howells
University of Melbourne, National Stroke Research Institute
AUSTRALIA
Background: Development of therapies for acute ischemic stroke frequently involves testing treatment efficacy in animal models of stroke, and evaluation in clinical trial. The amount of brain damage, or infarct volume, may offer a common metric to evaluate the ability of animal models to predict the clinical utility of a treatment. Method: Systematic review was used to identify experiments and clinical trials of acute stroke treatments. Controlled studies using infarct volume as an endpoint were selected for inclusion in the meta-analysis. Further analyses were performed to identify relationships between treatment efficacy and the type of stroke (e.g. size, location and cause), the treatment (e.g. time and mode of delivery) and the outcome (e.g. time and type of measurement). Results: 2932 experimental studies (n= 43,283 animals) and 40 clinical trials (n=5567 patients) were included in the analysis. Whilst treatments reduced infarct volume on average by 26% in animal models, and by 18-26% in clinical trials, there was less than perfect correspondence between clinical and animal findings for individual drugs. Nevertheless, both animal and clinical data support the notion that early drug delivery and early reperfusion contribute to a higher level of treatment efficacy, and that large as well as small strokes can be responsive to therapy. The nature of the control group was fundament to the interpretation of results: In clinical trials in particular, the baseline level of damage and choice of historical or placebo control was critical in the determination of outcome. Discussion: More work is needed to make animal studies commensurate with clinical trials, and to understand how infarct volume relates to other endpoints.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
09
An Ancrod Dosing Regimen for Acute Ischemic Stroke Derived from the Stroke Treatment with Ancrod Trial (STAT) and Validated in the European STAT (ESTAT) Trial
D.E.Levy
W.W.Wasiewski
E.B.Ringelstein
M. Kaste
H.C.Diener
G.J.del Zoppo
A.M.Pancioli
A.M.Demchuk
G. Howard
B. M.Demaerschalk
and G. W. Albers (not a trialist group)
Neurobiological Technologies, Inc.
USA
Several clinical trials of the defibrinogenating agent, ancrod, in acute ischemic stroke have been published, including the positive STAT (JAMA 2000) and A-20 trials (Stroke 1994) and more recently the negative ESTAT trial (Lancet 2006). Retrospective analysis of STAT showed that 52% of ancrod subjects with 0-3-hour defibrinogenation rates of at least 30 mg/dL/hr achieved efficacy based on a dichotomized Barthel Index vs. 26% of ancrod subjects with slower rates. Subjects with mean 9-hour-to-end-of-treatment (9H-EOT, planned: 5 days) fibrinogen levels above 60 mg/dL had a 0.8% incidence of symptomatic intracranial hemorrhage (SICH) vs. 21% with lower levels. Further, there were no SICHs in 220 ancrod subjects from STAT and A-20 with mean 9H-EOT fibrinogen levels above 70 mg/dL. These observations led to a hypothesis that a modified dosing regimen designed to achieve rapid initial defibrinogenation and avoid prolonged hypofibrinogenemia might be both effective and safe. The initial infusion rate in both STAT and ESTAT was based empirically on the pretreatment fibrinogen level. Subjects with levels of at least 450 mg/dL received the fastest initial ancrod infusion, 0.167 IU/kg/hr, and would be expected to have the fastest initial defibrinogenation. Efficacy was achieved by 44% of STAT ancrod subjects with this initial infusion rate vs. 25% of similarly-infused placebo patients; in the larger ESTAT population with that initial infusion rate, the difference (34% vs. 22%) was statistically significant. Confirming the findings in STAT and A-20, ESTAT ancrod subjects with mean 9H-EOT fibrinogen levels greater than 70 mg/dL had a 4% incidence of SICH compared to 16% with lower levels. While all of these analyses are retrospective, the validation in the failed ESTAT trial of a hypothesis generated in STAT has led to implementation of a modified dosing regimen in the ongoing Ancrod Stroke Program (ASP), in which subjects receive a single 2-3-hour infusion of placebo or ancrod at a rate of 0.167 IU/kg/hr. Whether this proves to be effective and safe requires completion of the two ASP studies.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
10
Calculation of sample size for stroke trials assessing functional outcome: comparison of binary and ordinal approaches
L.J.Gray
T. Collier
P.M.W.Bath
on behalf of the OAST Collaboration
Institute of Neuroscience, University of Nottingham
UNITED KINGDOM
Introduction: Many acute stroke trials have given neutral results. Sub-optimal statistical analyses may be failing to detect efficacy. Methods which take account of the ordinal nature of functional outcome data are more efficient. We compare sample size calculations for dichotomous and ordinal outcomes for use in stroke trials. Methods: Data from stroke trials studying the effects of interventions known to positively or negatively alter functional outcome - Rankin Scale and Barthel Index - were assessed. Sample size was calculated using comparisons of proportions, means, medians (according to Payne), and ordinal data (according to Whitehead). The sample sizes gained from each method were compared using Friedman 2 way ANOVA. Results: 55 comparisons (54,173 patients) of active versus control treatment were assessed. Estimated sample sizes differed significantly depending on the method of calculation (p<0.0001). The ordering of the methods showed that the ordinal method of Whitehead and comparison of means produced significantly lower sample sizes than the other methods. The ordinal data method on average reduced sample size by 28% (inter-quartile range 14% to 53%) compared to the comparison of proportions. The comparison of medians method gave the largest sample sizes. However, a 22% increase in sample size was seen with the ordinal method for trials assessing thrombolysis. Conclusions: Choosing an ordinal rather than binary method of analysis allows most trials to be, on average, smaller by approximately 28% for a given statistical power. Smaller trial sample sizes may help by reducing time to completion, complexity, and financial expense. However, ordinal methods may not be optimal for interventions which both improve functional outcome, and cause hazard in a subset of patients, e.g. thrombolysis.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
11
Post-stroke fatigue – a systematic review of its descriptions, measurements and co-existing clinical factors
A. Lerdal
L.N.Bakken
S. Eriksen
G. Pedersen
M. Kirkevold
A. Finset
K.A.Lee
H.S.Kim
Buskerud University Collge
NORWAY
Background Although fatigue is a common complaint after stroke, relatively little is known about how post-stroke fatigue is experienced and what its related factors are. Post-stroke fatigue as well as other symptom experiences of stroke requires an in-depth understanding in order to develop post-stroke rehabilitation programs. This review was undertaken to provide a comprehensive synthesis of knowledge from the literature concerning the description and definition and measurement of fatigue, its relation to stroke characteristics, socio-demographic factors and other clinical factors. Methods A search in PubMed, Cinahl and PsychLit was performed. “Cerebrovascular Accident” (in Pubmed and PsychLit) and “Cerebral Vascular Accident” (in Cinahl) were used as Medical Subject Headings in combination with “fatigue” as a key word in all three databases. Results Descriptions of fatigue revealed multiple dimensions of the phenomenon. No specific, unique definition was found of fatigue as a post-stroke phenomenon. The most frequently used instruments for post-stroke fatigue were the Fatigue Severity Scale and a Fatigue Visual Analogue Scale. None of the scales used in the stroke populations have been developed specifically for measuring post-stroke fatigue. Age, gender, living conditions, and personality showed to have some associations with post-stroke fatigue, albeit conflicting. Studies showed a discrepancy in findings with regards to the relationship between stroke-related characteristics such as stroke location/type, the numbers of stroke, neurological deficits and fatigue. There is an indication for an association between pre-stroke and post-stroke fatigue. Discussion The review shows that our knowledge regarding this phenomenon is still limited. There is a need to continue empirical research with various theoretical orientations. Possible antecedent components identified are personal factors, biomarkers, stroke characteristics, pre-stroke fatigue, and co-morbid diseases.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
12
Is providing informal care to stroke survivors associated with low mood? A systematic review and meta-analysis
L.A.Legg
A. Falconer
C. Weir
D.J.SStott
L.N.Smith
P. Langhorne
University of Glasgow
UNITED KINGDOM
Background: Stroke is a major cause of adult disability and at any one time 0.5% of the UK population are dependent as the result of a stroke. Much of the burden of providing for the needs of these people falls to an informal caregiver. Objective: To determine whether there is an increased risk of low mood among those exposed to care giving and whether there is evidence of a causal relationship between exposure to care giving and low mood. Methods: We searched for published or unpublished longitudinal studies of care giving exposure, written in any language that recruited participants prospectively during the care giving period and used a control or comparison group. Two reviewers independently selected studies for eligibility. The standardized mean difference from a random effects model was used to estimate the association between caregiving exposure and low mood, taking account of statistical heterogeneity. Results: We included five studies (724 participants); assessing mood using SF-36 (3 studies), Short Zung Depression scale and Beck Depression Inventory. Caregivers had lower mood than non-caregivers at interim (median 6 months) follow-up (standardised mean difference –0.29; 95% CI -0.40 to -0.18; p<0.00001) and at end of scheduled (median 12 months) follow-up (-0.32 (-0.41 to -0.22; p<0.00001) with no significant heterogeneity. Conclusion: This meta-analysis provides an appraisal of the studies most likely to provide a reliable estimate of the adverse health effects of providing care to stroke survivors. Caregivers are more likely to have low mood than non-caregivers. A major drawback in these studies is that caregiving precedes initial measurement of depression, making it impossible to conclude that the observed association is a causal one.
Session:
Poster Session I
Date:
Wednesday 14 May 2008
Time:
12:30 - 14:00
Room:
Agora 3
13
Animal Models of Lacunar Stroke? – A Systematic Review
E.L.Bailey
J.M.Wardlaw
J. McCulloch
C. Sudlow
University of Edinburgh
UNITED KINGDOM
Background: Lacunar ischaemic stroke accounts for up to 25% of all ischaemic strokes. It is caused by small subcortical infarcts in the white and deep grey matter. The origin of the small vessel pathology leading to these lesions is unknown. Possible explanations include thromboembolic occlusion or endothelial dysfunction. Experimental models are required to evaluate treatments for lacunar stroke and help to explain the human pathophysiology. Methods: We systematically reviewed the published literature on animal models which could represent lacunar stroke. Studies were identified with a validated search strategy across 3 databases, along with hand searching of two key journals. We included studies that could represent an aspect of lacunar stroke, as well as those aiming to model a condition with similar pathology (e.g. vascular dementia). Results: From a total of 5670 papers, 40 studies (45 papers) met inclusion criteria, representing over 10 different induction techniques: e.g. emboli, endothelin-1 injections and spontaneous lesions in aged or genetically modified animals. Meta analysis could not be performed as data on infarct size was too diverse and sample size was often not stated. Most studies focussed on producing ischaemic lesions in brain tissue, with only one (the Spontaneously-Hypertensive Stroke Prone Rat) addressing the small vessel pathology - in this model the mechanism seems to concern endothelial dysfunction rather than vessel occlusion. In general, infarcts produced in animals were too large or affected the wrong brain territory to be pertinant to lacunar stroke. Embolic models indicate that most lacunar stroke is unlikely to be due to an embolic source. Discussion: Some currently available animal models replicate small deep infarcts and could be used to evaluate new treatments to reduce ischaemic damage, but mostly do not mimic human small vessel pathology so are a limited model of human disease. Improved models of small vessel disease are needed to focus on the pathological events that cause most lacunar stroke.
