Abstract WMP67: Stroke Screening Tools Have High Specificity for Detecting Large Vessel Occlusion in a Southeastern US Prospective Cohort Study

Introduction: Previously reported stroke screening tools used to identify large vessel occlusion (LVO) have been developed based on retrospective analyses or cohorts enrolled in a clinical trial. Hypothesis: Stroke screening tools may have lower specificity for detection of LVO when tested in a prospective cohort study of patients identified by emergency medical services (EMS) to have stroke signs and symptoms. Methods: Consecutive patients presenting to a stroke center with stroke symptoms or a positive FAST screen by EMS within 6 hours from stroke onset or wakeup were included in this prospective analysis. All patients were initially evaluated by a neurology attending or resident and underwent non-contrast CT (NCCT). Patients identified to have a stroke mimic on initial assessment or hemorrhagic stroke on NCCT did not undergo intracranial large artery imaging (iLAI); all remaining patients underwent iLAI. Sensitivity and specificity for LVO with the following screens were assessed: NIHSS ≥ 15, LAMS ≥ 4, RACE ≥ 5 and CPSSS ≥2. Results: Of 92 patients (mean age 69±17 years, mean NIHSS 8) evaluated over the 3 month study period, 41 (45%) were identified to have a stroke mimic and 9 (10%) had a hemorrhagic stroke on NCCT; 19 (21%) patients had LVO. The remaining 42 patients underwent iLAI within Conclusion: Previously published stroke screening tools have high specificity for detection of LVO when tested in a prospective cohort study of patients identified by EMS to have stroke signs and symptoms.