Abstract TMP64: Single Stroke Symptoms Cannot Predict Large Vessel Occlusion in Patients With Acute Ischemic Stroke With a High Reliability

Abstract

Background: Identification of acute ischemic stroke patients (AIS) with large vessel occlusion (LVO) in the pre-hospital setting is paramount now that endovascular treatment has become the standard of care. Correctly identifying this AIS population will allow transport to appropriate stroke centers with endovascular stroke capabilities. Objectives: We choose to evaluate single symptoms which, if present, might signify an ischemic stroke with a LVO. We tested if the following stroke symptoms including hemiplegia, severe dysarthria, neglect, eye deviation were associated with a LVO. Our hypothesis is that the use of these selected single or multiple NIHSS symptoms will not result in a ROC of > 0.9 rendering it effective in LVO diagnosis. Methods: A retrospective cohort review was conducted of 551 AIS patients who presented to the emergency department over a 77 month period. Subject demographics clinical symptoms, and other pertinent information were collected from the medical record. The following symptoms at ED presentation were evaluated with univariate and multivariate regression and ROC curve analysis; hemiplegia (of arm and leg), dysarthria (severe), neglect, and eye deviation. Results: Out of the 551 patients, 220 (40%) had a moderate stroke (NIHSS >9) while 143 (26%) had a severe stroke (NIHSS>14) and 49% were male. The median NIHSS was 7 (IQR 4-15) and 228 (41%) had LVO on CTA. The odds ratio and area under the ROC curve for hemiplegia, severe dysarthria, neglect, and eye deviation were respectively OR 3.2 (2.2- 4.5), 3.2 (2.2- 4.6), 5.5 (3.4- 8.9), 7.7 (4.6- 13.1) and area under the ROC 0.63, 0.64, 0.63, and 0.64 respectively. The results of multivariate analysis were slightly better with an ROC of 0.72. Conclusions: Singular stroke symptoms do not have the diagnostic power to predict which patient with AIS will have a LVO. While they have positive odds ratios, the ROC analysis suggests that individually or in combination, they are not accurate enough to determine the presence of LVO.