Abstract 82: Endovascular Therapy Eligibility in the Advent of DAWN and DEFUSE-3 Results

Abstract

Introduction: We previously estimated that, among 2297 ischemic stroke patients in our 1.3 million metropolitan population in 2010, 159 were potentially IV rtPA (IVT)- and 29 were endovascular (EVT)-eligible based on AHA Class I, Level A guideline criteria; this projected to 31,000 IVT and 5600 EVT-eligible patients in the US annually. Given recent DAWN and DEFUSE 3 results suggesting EVT benefit beyond 6 hours to 24 hours, we aim to estimate the number of additional EVT-eligible patients in our region. Methods: We ascertained all hospitalized AIS patients ≥18 years old in 2010 using ICD-9 codes 430-436 within Greater Cincinnati/Northern Kentucky (GCNK) population; all cases were physician-reviewed. GCNK patients were considered potentially DAWN-eligible if ≥18 years, presenting within 5 to 23 hours from onset, and with a baseline NIHSS ≥10, prestroke mRS 0-1, platelets ≥50k, glucose 51-399 mg/dL, INR ≤3.0 and creatinine ≤3.0. Based on single-center (Pittsburgh) data, an eligibility proportion of 22% for both large vessel occlusion (LVO) and clinical-imaging-mismatch (CIM) was then applied. Additional patients eligible by only DEFUSE 3 criteria were identified within the GCNK population by including NIHSS 6-9, prestroke mRS 2, and an upper age of 90 years among patients presenting within 5 to 15 hours. Based on single-center (Bern) data, LVO rates by NIHSS score were applied. Based on DEFUSE 3 screening estimates, a 50% target mismatch (TTM) rate was then applied. Results: Among 2297 ischemic stroke patients hospitalized in our region, 34 were potentially DAWN-eligible, including 8 estimated to meet LVO and CIM criteria, and 19 additional patients were DEFUSE 3-eligible, including 4 estimated to meet LVO and TMM criteria. Conclusion: The 6-24 hour EVT treatment window may increase EVT eligibility by approximately 40% in the GCNK metropolitan region annually, although this estimate is limited by small sample sizes and inferences. Estimates will be updated as further data become available by the time of presentation of this abstract.