Abstract WP248: Comparison of Large Vessel Occlusion Prediction Scores

Background: There is an urgent need to rapidly identify patients with acute ischemic stroke due to large vessel occlusion (LVO) to improve access to endovascular therapy. Aim: To derive a scale to improve pre-hospital detection of LVO. Methods: A retrospective review of consecutive patents enrolled in our prospective Comprehensive Stroke Center registry from 2008-2013 was conducted. Patients were excluded if they were < 18 yrs, not ambulatory, or did not have MRA or CTA. Demographic and clinical characteristics were compared between patients with and without LVO, defined as occlusion of the intracranial internal carotid artery, basilar artery, or middle cerebral artery. We compared a novel Large Vessel Occlusion Scale (LVOS) to the Cincinnati Prehospital Stroke Severity Scale (CPSSS) by constructing ROC curves and determining their area under the curve (AUC). Results: A total of 1,663 patients were identified (mean age = 62 yrs, SD = 14, 45.6% women and 66.8% Black). LVO was present in 171 patients (10.3%). The LVOS is a 6 point scale derived from the NIHSS that includes 2 points for abnormal speech or language (NIHSS ≥ 1 for either), 1 point for gaze preference (NIHSS ≥ 1), 1 point for facial asymmetry (NIHSS ≥ 1), 1 point for asymmetrical arm drift (RUE ≠ LUE), and 1 point for asymmetrical leg drift (RLE≠LLE). The LVOS had an AUC of 0.688 (95% CI 0.736 - 0.640), compared to 0.646 (95% CI 0.693 - 0.598) for the CPSSS and 0.678 (95% CI 0.0.723 - 0.633) for the NIHSS. A LVOS score ≥ 3 was 75.4% sensitive (95% CI 68.5 - 81.4) and 50.3% specific, positive likelihood ratio of 1.517 and negative likelihood ratio of 0.162 in predicting LVO. A CPSSS score ≥ 2 was 46.8% sensitive (95% CI 39.7 - 54.0) and 90.0% specific, positive likelihood ratio of 2.870 and negative likelihood ratio of 0.238 in predicting LVO. Conclusions: The LVOS compares favorably to the CPSSS in its ability to identify patients with LVO. Prospective prehospital validation is needed.