XVII. European Stroke Conference
Nice, France

Oral Session:

Prognostic predictors
Date: Wednesday 14 May 2008  
Time: 16:30 - 16:40 -  Room: Euterpe
Chair: D. Tanne, Israel and E. Touze, France

01
Admission hyperglycaemia predicts a worse outcome in a large community-based cohort of acute ischaemic stroke patients treated with intravenous thrombolysis
A.Y.Poppe    S.R.Majumdar    T. Jeerakathil    W. Ghali    A.M.Buchan    M.D.Hill                            
on behalf of the CASES Investigators

Calgary Stroke Program, Foothills Medical Centre University of Calgary

CANADA

Background: Admission hyperglycemia has been associated with worse outcomes in ischemic stroke patients, including those treated with intravenous tPA (IV-tPA). We hypothesized that admission hyperglycemia (glucose >8.0 mmol/L) would be independently associated with increased rates of mortality, symptomatic intracerebral hemorrhage (SICH) and poor functional status at 90-days in a large community-based cohort of stroke patients treated with IV-tPA. Methods: Using data from the Canadian Alteplase for Stroke Effectiveness Study (CASES), a national, prospective registry of stroke patients treated with IV-tPA, the association between admission hyperglycemia and prespecified outcomes was examined. Multivariable logistic regression was used to determine if admission hyperglycemia was independently associated with mortality, SICH, and poor functional status at 90-days (modified Rankin Scale [mRS] >1). Similar analyses were conducted examining glucose as a continuous measure. Results: Of 1098 patients, median age was 73 years, 45% were women, 96% had anterior circulation stroke, and 16% had diabetes mellitus. Overall, 296 (27%) had admission hyperglycemia, including 18% of those without diabetes and 70% of those with diabetes. Admission hyperglycemia was associated with increased risk of death (RR 1.64, 95%CI 1.31 to 2.06), SICH (RR 1.87, 95%CI 1.07 to 3.25) and decreased probability of a favourable outcome (mRS 0-1) at 90 days (RR 0.69, 95%CI 0.56 to 0.85). With glucose as a continuous variable, an incremental risk of death, SICH and unfavourable 90-day outcomes was observed with increasing admission glucose values. These relationships held true for patients with and without diabetes. Discussion: In this representative cohort of acute stroke patients treated with IV-tPA, admission hyperglycemia was independently associated with an increased risk of death, SICH and poor functional status. Irrespective of whether better glycemic control can improve these outcomes, admission hyperglycemia identifies a population at high risk of poor outcomes.

Prognostic predictors
Date: Wednesday 14 May 2008  
Time: 16:40 - 16:50 -  Room: Euterpe
Chair: D. Tanne, Israel and E. Touze, France

02
Blood biomarkers to improve prediction of the prognosis in ischaemic stroke: A systematic review
W.N.Whiteley    W.L.Chong    A. Sengupta    P. Sandercock                                          
 

Division of Clinical Neurosciences, University of Edinburgh

UNITED KINGDOM

INTRODUCTION Blood biomarkers could add predictive ability to clinical prognostic models. We sought to review the literature systematically. METHODS We searched Medline and EMBASE from 1966 to 3rd January 2008, reference lists and personal files for all studies measuring a blood marker at admission and recording its relationship to a clinical outcome after ischaemic stroke. 3 authors reviewed the papers and extracted data, resolving differences by discussion. RESULTS We identified 6020 studies. 81 studies of 69 blood markers were examined. 61/81 studies did not report blinding of marker measurement to stroke status, 20/81 did not prespecify poor clinical outcome, 18/81 did not give an adequate description of marker measurement and 52/81 did not examine unselected stroke patients. There was no adjustment for age or stroke severity in 35/81 studies. In 48/81 studies the first blood samples were drawn <24 hours after stroke onset. Analysis of the association between biomarker level and poor outcome was significant (p<0.05) for the following proportions of markers of: fibrinolysis 3/8 (38%); anti-inflammatory 2/12 (17%); cardiac dysfunction 9/12 (75%); cell damage 2/2 (100%); endothelium 5/9 (56%); glial cells 5/9 (56%); neuronal damage 3/8 (38%); haemostasis 18/42 (43%); inflammation 27/53 (51%); and neurotransmitters 4/4 (50%). CONCLUSIONS. No class of marker, other than those of cardiac damage, was reasonably consistently associated with poor outcome. Methods were weak: none assessed if the biomarker added predictive ability to a clinical model, few were blinded & cohorts were no ideal. Future studies should: prespecify outcomes, blind measurement of biomarker and outcome, examine unselected cohorts of stroke patients & assess if biomarkers add power to validated clinical models.

Prognostic predictors
Date: Wednesday 14 May 2008  
Time: 16:50 - 17:00 -  Room: Euterpe
Chair: D. Tanne, Israel and E. Touze, France

03
Are Shape Irregularity and Density Heterogeneity on CT Predictors of Hematoma Growth in Intracerebral Hemorrhage?
C.D.Barras    S. Christensen    L. MacGregor    B.M.Tress    D.A.De Silva    S.M.Davis                            
for the Recombinant Activated Factor VII Intracerebral Hemorrhage Trial Investigators

Royal Melbourne Hospital, The University of Melbourne

AUSTRALIA

Background: Intracerebral hemorrhage (ICH) growth is a determinant of mortality and functional outcome. We hypothesised that irregular shape and heterogeneous density are predictive of ICH growth. The aims were to determine if: i) ICH shape or density predict growth and if ii) ICH size affects shape or density. Methods: Blinded to 24hr growth data, 3 raters assessed baseline CTs of 90 placebo group ICH patients from a hemostatic agent trial, obtained within 3hrs of ictus. Raters independently applied novel 5-point categorical scales of shape: category (cat.) 1-5 (most regular to most irregular) and density: cat. 1-5 (most homogeneous to most heterogeneous). Baseline ICH volumes were divided into tertiles: small (0-10mL), medium (>10-25mL) and large (>25-106mL). Results: Rater agreements exceeded 85% (+/- 1 category) for both categorical scales. Baseline median ICH volume was 14mL (IQR 28mL). Median volume change was 2.5mL (IQR 11.5mL). Median time to baseline scan was 1.75 hrs (IQR 0.63hrs) from ictus. Median growth was significantly higher in large (7.39mL, IQR 17.8) compared to small (1.10mL, IQR 2.08) ICHs (p<0.001). Median growth was significantly less in homogeneous (cat 1-2) (1.21mL, IQR 3.77) compared to heterogeneous (cat 3-5) ICHs (5.07mL, IQR 18.01) (p=0.008). Median growth was higher in irregular (cat 3-5) (2.69mL, IQR 13.14), compared to regular (cat 1-2) ICHs (1.32mL, IQR 4.51) trending towards significance (p=0.084). Small ICHs were more regularly shaped (43%) than the medium (17%) and large (3%) ICHs (p<0.001). Small ICHs were more homogeneous (73%) compared to medium (37%) and large (17%) ICHs (p<0.001). Regardless of growth model, after adjustment for baseline ICH volume and time to scan, neither shape nor density independently predicted ICH growth. Discussion: Novel reliable categorical shape and density heterogeneity scales were developed. Large ICHs were significantly more irregular in shape, heterogeneous in density, and grew more. Shape irregularity and density heterogeneity are features, but not independently predictive, of ICH growth.

Prognostic predictors
Date: Wednesday 14 May 2008  
Time: 17:00 - 17:10 -  Room: Euterpe
Chair: D. Tanne, Israel and E. Touze, France

04
PAI-1 4G5G polymorphism is associated with brain vessel reocclusion rates after fibrinolytic therapy among ischemic stroke patients.
I. Fernandez-Cadenas    M. Rubiera    M. Mendioroz    A. Del Rio    S. Domingues-Montanari    M. Ribo    C.A.Molina    A. Rosell    J. Alvarez-Sabin    J. Montaner
 

Neurovascular Research Laboratory and Neurovascular Unit, Institut de Recerca, H. Vall d'Hebron.

SPAIN

Background: Despite t-PA proven benefits, up to 13% of stroke patients suffer reocclusions after t-PA administration thus the obtained clinical benefit is vanished. We aimed to analyse whether functional polymorphisms in fibrinolysis inhibitor genes [Plasminogen Activator Inhibitor 1 (PAI-1) and thrombin activatable fibrinolysis inhibitor (TAFI)] might be associated with reocclusion rates after thrombolytic therapy. Methods: 165 patients with ischemic stroke involving the MCA territory who received t-PA <3h were studied. PAI-1 4G5G polymorphism determination was performed by direct sequencing and TAFI Thr325Ile polymorphism by RFLP. PAI-1 mRNA was studied by Real-Time PCR analysis. Recanalization and reocclusion was diagnosed by means of Transcranial Doppler. NIHSS was serially measured to assess neurological outcome and modified Ranking Scale (mRS) at 3rd month was used to evaluate functional outcome. Results: Although no association was found between TAFI genotype and reocclusion rates, PAI-1 4G4G patients had higher reocclusion rates (4G4G= 12.5% vs other genotypes= 2.7%, p=0.025). In a logistic regression, 4G4G genotype was the only factor associated with reocclusion (OR=15.16 95% CI 1.4–163.4, p=0.025). 4G4G genotype was also associated with poor functional outcome at 3rd month (4G4G= mRS 4 vs others genotypes= mRS 3; P=0.017) and with PAI-1 mRNA levels at 12 hours after symptoms onset (Patients 4G4G: 2.01% versus other genotypes: 0.68%; P=0.034). Discussion: PAI-1 4G4G genotype was associated with reocclusion rates and poor functional outcome among stroke patients treated with t-PA. The transcriptional functionality of this polymorphism suggests that blocking PAI-1 pathway might increase sustained brain reperfusion.

Prognostic predictors
Date: Wednesday 14 May 2008  
Time: 17:10 - 17:20 -  Room: Euterpe
Chair: D. Tanne, Israel and E. Touze, France

05
Low ankle-brachial index (ABI) predicts cardiovascular risk after acute cerebral ischemia
M.A.Busch    K. Lutz    J.E.Röhl    F. Masuhr                                          
 

Charite - University Medicine Berlin

GERMANY

BACKGROUND: Growing evidence supports the association between a low ankle-brachial blood pressure index (ABI) and an increased risk of death and cardiovascular disease in the general population and in high-risk individuals. Little is known, however, about the prognostic value of a low ABI in patients with acute cerebral ischaemia. We aimed to determine whether a low ABI predicts future risk of stroke, myocardial infarction (MI) and death after an ischaemic stroke or transient ischaemic attack (TIA). METHODS: We prospectively studied 197 consecutive hospital-admitted patients with acute ischaemic stroke or TIA who underwent ABI measurement at baseline and were followed for up to 4 years. In line with previous studies, an ABI <=0.9 was defined as low. Primary outcome was the composite of stroke, MI or death. Association between low ABI and time to first outcome event was examined by survival analysis and by Cox regression analysis adjusted for potential confounders. RESULTS: Among 197 included patients (44% female, median age 64yrs, 25% TIA), 61 (31%) had a low ABI. Older age, smoking, hypertension and hyperlipidaemia were associated with a low ABI. During a median follow-up of 2.3 years, 58 outcome events (34 IS, 4 MI, 20 deaths) occurred in 38 patients (19%). There were 453 person-years available for survival analysis. In age- and sex-adjusted survival analysis, a low ABI was associated with a more than doubled risk of an outcome event (hazard ratio (HR), 2.17; 95% CI 1.10-4.34; p=0.02). This association remained unchanged after further adjustment for risk factors and symptomatic arteriosclerosis. CONCLUSIONS: A low ABI was found in a third of patients and was an independent risk marker for cardiovascular events and death after an ischaemic stroke or TIA.

Prognostic predictors
Date: Wednesday 14 May 2008  
Time: 17:20 - 17:30 -  Room: Euterpe
Chair: D. Tanne, Israel and E. Touze, France

06
High resolution MRI for detection of high risk plaques in moderate carotid artery stenosis
L. Esposito    P. Heider    O. Wolf    R. Feurer    K. Holzer    S. Sadikovic    D. Sepp    T. Liebig    B. Hemmer    H. Poppert
 

Department of Neurology; Technical University of Munich, Germany

GERMANY

Background Invasive therapy reduces the risk of stroke in high-grade (>70%) carotid artery stenosis (CAS). As the benefit of invasive therapy in moderate CAS (50-70%) is discussed controversially, new approaches for risk stratification in these patients are required. High-Resolution Magnetic Resonance Imaging (MRI) represents a promising tool for risk classification of carotid plaques. MRI-Plaqueimaging allows categorization in distinct lesion types (I-VIII) according to a modified histological scheme based on American Heart Association (AHA) criteria. Lesion types IV-V and VI are regarded as high risk plaques. This study evaluates weather MRI is able to reveal high risk carotid plaques in patients presenting with symptomatic and asymptomatic moderate CAS. Methods 35 patients (15 female) with moderate (50-70%) CAS were enrolled and imaged using a 1.5 T scanner with bilateral phased-array carotid coils. T1-, T2-, time-of-flight (TOF) and proton-density (PD)-weighted studies were obtained. The carotid plaques were classified as lesion types I–VIII according to the MRI-modified AHA criteria. Results 10 patients presented with a recently (< 3 days) symptomatic CAS. One patient had to be excluded due to insufficient MR-image quality. In symptomatic patients MRI-detected unstable lesion types were predominant as compared to asymptomatic controls (60% vs. 16 %; P=0.032; two-tailed Pearson-Test). Discussion In patients presenting with a symptomatic CAS MRI-detected high risk lesions were predominant. MRI-based plaque-imaging seems to represent a promising tool for the detection of unstable carotid plaques and for risk stratification of patients presenting with moderate CAS. MRI could improve the selection of patients presenting with a symptomatic CAS for invasive therapy.

Prognostic predictors
Date: Wednesday 14 May 2008  
Time: 17:30 - 17:40 -  Room: Euterpe
Chair: D. Tanne, Israel and E. Touze, France

07
Female gender is a predictor of worse outcome in patients treated with thrombolysis for acute stroke
P. Martinez-Sanchez    B. Fuentes    M. Alonso de Leciñana    J. Masjuan    J. Egido    P. Simal    F. Diaz-Otero    A. Gil-Nuñez    E. Diez-Tejedor       
 

Hospital Universitario La Paz. Madrid

SPAIN

BACKGROUND: there are few studies analysing sex differences in outcome after thrombolitic treatment for acute ischemic stroke and they show contradictory results. METHODS: multicentric observational study with acute stroke patients treated with endovenous rtPA in three university hospitals (2003-2006). We analysed the following variables by gender: vascular risk factors, stroke severity (NIHSS), glycaemia and systolic blood pressure (SBP) previous to rtPA infusion; etiological stroke subtype, in-hospital complication (haemorrhagic transformation, brain oedema), mortality and 3-moths outcome by the modified Rankin Scale (mRS) score. RESULTS: 325 patients, 144 women. Mean age 65.6 years (SD 13.3). Women had a lower onset to treatment time (P = 0.018), lower frequency of smoking (P<0.0001) as well as arterial origin strokes (P = 0.001), but higher frequency of atrial fibrillation (P = 0.007) and cardioembolic strokes (P = 0.008) than men. At 3-months, the 44.2% of women and the 28.8% of men had a mRS score >3 (P = 0.008). The logistic regression analysis showed that female (OR, 2.048; 95% CI, 1.027 to 4.081) was a predictor of 3-moths poor outcome; after adjustment for age, sex, previous vascular risk factors, glycaemia, SPB, stroke subtype, stroke severity and in-hospital complications. DISCUSSION: female gender is a predictor of poor outcome after thrombolytic treatment for acute ischemic stroke.

Prognostic predictors
Date: Wednesday 14 May 2008  
Time: 17:40 - 17:50 -  Room: Euterpe
Chair: D. Tanne, Israel and E. Touze, France

08
Baseline diffusion-weighted imaging lesion extent: a potential marker of outcome in acute basilar artery occlusion
T.H.Cho    N. Nighoghossian    F. Tahon    C. Némoz    M. Hermier    F. Salkine    L. Derex    S. Cakmak    P. Trouillas    F. Turjman
 

Neurovascular emergency department, Hôpital Neurologique, Lyon 1 University

FRANCE

Background Vertebrobasilar occlusion (VBO) is the most severe ischemic stroke subtype with mortality reaching 80-90% without recanalization. Recanalization is the strongest predictor for outcome. Acute magnetic resonance imaging (MRI) may be of additional prognostic value. Our purpose was to assess the impact of pretreatment diffusion-weighted imaging (DWI) lesion extent on clinical outcome. Methods We analyzed in retrospect the baseline clinical and MRI data of patients treated with endovascular procedures for acute VBO. Age, National Institutes of Health Stroke Scale (NIHSS), Glasgow Coma Scale, duration of coma and time to treatment were analyzed. Outcome was assessed by the modified Rankin Scale (mRS) and classified as favorable (mRS 0-2) or unfavorable (mRS 3-6). Brainstem damage was assessed by the total number of arterial territories with abnormal DWI in the medulla, pons and midbrain (brainstem DWI score). Cerebellar lesions were dichotomized as minor or major (less or more than a third of cerebellar hemisphere, respectively). Univariate and multivariate regression analysis were used to identify clinical and MRI predictors of outcome. Results Thirty-five consecutive patients (14 women and 21 men, mean+/-SD age, 56+/-14.4 years) were included. Median NIHSS score was 22 (range, 4-34). Mean+/-SD time to treatment was 554.8±320.3 minutes. Median brainstem DWI score was 3 (range, 0-14). Major cerebellar lesions were observed in 41% of patients. Recanalization was achieved in 26/35 patients. Without recanalization, none achieved a favorable outcome. Higher baseline NIHSS score (p=0.005), coma at admission (p=0.007), protracted coma (p=0.006), higher brainstem DWI score (p=0.004) and major cerebellar damage (p=0.009) were significantly associated with an unfavorable outcome. Multivariate regression analysis showed that the brainstem DWI score was the strongest independent pretreatment predictor of clinical outcome (p=0.026). Conclusion Pretreatment brainstem DWI lesion extent is a strong predictor of clinical outcome for recanalized patients.

Prognostic predictors
Date: Wednesday 14 May 2008  
Time: 17:50 - 18:00 -  Room: Euterpe
Chair: D. Tanne, Israel and E. Touze, France

09
Characteristics and short term prognostic factors in 300 stroke patients aged 80 and over
F. Dumont    M. Rutgers    C. Lefebvre    E. Janssens    D. Leys    H. Henon                            
 

Service de Neurologie B. Neurovasculaire. Hopital Roger Salengro. CHRU de Lille

FRANCE

Introduction: The prevalence of very old stroke patients is rising. However, this population is poorly characterised. Objectives: To define stroke characteristics and factors influencing short term outcome in stroke patients aged 80 years and over. Method: 300 consecutive patients ≥ 80 years admitted for a stroke or a TIA in the neurology units of Lille University Hospital were included. Demographics, medical history, stroke risk factors, previous functional and cognitive status and initial stroke severity were recorded. Stroke etiology was determined according to the TOAST criteria. Vital and functional outcome were assessed at discharge. Results: The median age was 84 years, 112 patients were males. 230 (77%) patients were independent before stroke. 268 (89%) were admitted for ischemic stroke (53 TIA), 32 (11%) for hemorrhage. In ischemic stroke patients, the most frequent etiologies were cardioembolism (36,6%) and undetermined due to incomplete diagnostic work-up (31,3%). 130 patients (43,3%) experienced a severe stroke (NIHSS score ≥ 6). Factors independently associated with initial stroke severity were: IQCODE score (OR 1,34; IC95% 1,003-1,066), atrial fibrillation (OR 3,156; IC95% 1,636-6,086) and early signs of stroke on admission CT scan (OR 2,214; IC95% 1,151-4,257). 45 patients died during hospitalisation (15%; IC95% 11-19). Factors independently associated with hospital death were: stroke severity (OR=1,132; IC95% 1,093-1,171) and history of stroke (OR 2,489; IC95% 1,247-4,967). 143 (56%) survivors were independent at discharge, stroke severity being the only predictive factor for independence (OR 0,825; IC95% 0,778-0,875). Conclusion: These data emphasise the importance of initial stroke severity on short term prognosis in very old stroke patients.