Page 11

Karger_ESC London_2013

London, United Kingdom 2013 4 Acute stroke: current treatment & new treatment concepts 9:00 - 9:10 Optimisation of sample size calculation in thrombectomy trials using the natural history of proximal occlusions after i.v. thrombolysis failure A.I. CALLEJA1, E. CORTIJO2, P. GARCÍA-BERMEJO3, C. DE LA CRUZ4, J REYES5, N. TELLEZ6, N FERNÁNDEZ-BUEY7, J.F ARENILLAS8 HOSPITAL CLÍNICO UNIVERSITARIO, VALLADOLID, SPAIN1,HOSPITAL CLÍNI-CO UNIVERSITARIO, VALLADOLID, SPAIN2, HOSPITAL CLÍNICO UNIVERSITARIO, VALLADOLID, SPAIN3, HOSPITAL CLÍNICO UNIVERSITARIO, VALLADOLID, SPAIN4, HOSPITAL CLÍNICO UNIVERSITARIO, VALLADOLID, SPAIN5, HOSPITAL CLÍNICO UNIVERSITARIO, VALLADOLID, SPAIN6, HOSPITAL CLÍNICO UNIVERSITARIO, VAL-LADOLID, SPAIN7, HOSPITAL CLÍNICO UNIVERSITARIO, VALLADOLID, SPAIN8 BACKGROUND:Stroke patients who do not respond to intravenous thrombolysis are being considered for endovascular rescue therapy worldwide. For many stroke physicians, random-ized controlled clinical trials(RCTs)to evaluate safety and efficacy of endovascular treatment represent an ethical problem. Ideally, such trials should be powered to provide rapidly definite response with minimum sample size needed. We applied our data on the natural history of this group of patients to estimate the minimum sample size needed to prove superiority of rescue endovascular-treatmentMETHODS:We included consecutive ischemic stroke patients with an acute intracranial occlusion in anterior circulation, with a baseline NIHSSscore >10, who were treated with intravenous-thrombolysis, and showed persistence of arterial occlusion after i.v.-therapy. A superiority clinical trial was designed with Ene 3.0 assuming an overall differ-ence of >10% in good outcome favouring endovascular therapy, drop-out rate 10%, alpha-error 2.5%, beta-error 10%. Outcome of i.v group was inferred from our sample of i.v failure, where-as a wide range of expected outcomes in the endovascular arm was extracted from main throm-bectomy registries with stent-retriever-devices. Primary outcome variable was modified Rankin scale </=2. RESULTS:We included 115 stroke patients with persistent intracranial arterial occlusions 1 h after i.v. thrombolysis. Poor outcome was observed in 83(71%)patients. A ROC-curve identi-fied a NIHSS cutoff-point of 16 that best discriminated outcome. Among patients with baseline NIHSS >16 and lack of early response to tPA, only 15% achieved good outcome. Assuming an intermediate response rate in the stenttriever arm of 45%, the sample size dramatically fell from 2,137 to 112 patients per arm when baseline NIHSS entry threshold was moved from initial 10 to 16. CONCLUSION: An increase of baseline NIHSS from 10 to 16 as inclusion criteria would dra-matically reduce the sample size needed to prove superiority of endovascular therapy in a RCT. Cerebrovasc Dis 2013; 35 (suppl 3)1-854 11 Table: Clinical variables associated with Major clinical improvement post-mechanical throm-bectomy. Minimal clini-cal improvement (<50% change in NIHSS from base-line) Major clinical im-provement (>50% change in NIHSS) P N 38 67 Median Age (IQR) 69.5 (20) 67(16) 0.408 Proportion Male gender 63.89% 44.83% 0.036* Proportion Hypertension (SBP>140) 53.57% 49.02% 0.349 Proportion Diabetes 21.43% 15.69% 0.261 Proportion Hyperlipidemia 48.15% 36.73% 0.166 Proportion Current Smoking 24.14% 18.75% 0.286 Proportion Atrial Fibrillation 25.0% 28.0% 0.6325 Proportion with Coronary artery disea-se 32.14% 24.49% 0.2342 Proportion Antiplatelet 45.83% 37.5% 0.249 Proportion Anticoagulation 21.74% 13.04% 0.176 Median baseline NIHSS(IQR) 17.5(7) 17 (8) 0.513 Proportion treated with IV-tPA 45.45% 70.21% 0.026* Median imaging to reperfusion (min) 116.5 (82.5) 94 (70) 0.029** Median puncture to reperfusion(min) 54 (53) 33 (29) 0.021** Median onset to reperfusion(min) 265 (140) 257(251.5) 0.794 * Two-sample test of proportions p>0.05 ** Kruskal-Wallis equality of proportions rank test p>0.05


Karger_ESC London_2013
To see the actual publication please follow the link above