Large Vessel Occlusion Scales Increase Delivery to Endovascular Centers Without Excessive Harm From Misclassifications

Abstract

Background and Purpose— Clinical large vessel occlusion (LVO) triage scales were developed to identify and bypass LVO to endovascular centers. However, there are concerns that scale misclassification of patients may cause excessive harm. We studied the settings where misclassifications were likely to occur and the consequences of these misclassifications in a representative stroke population. Methods— Prospective data were collected from consecutive ambulance-initiated stroke alerts at 2 stroke centers, with patients stratified into typical (LVO with predefined severe syndrome and non-LVO without) or atypical presentations (opposite situations). Five scales (Rapid Arterial Occlusion Evaluation [RACE], Los Angeles Motor Scale [LAMS], Field Assessment Stroke Triage for Emergency Destination [FAST-ED], Prehospital Acute Stroke Severity scale [PASS], and Cincinnati Prehospital Stroke Severity Scale [CPSSS]) were derived from the baseline National Institutes of Health Stroke Scale scored by doctors and analyzed for diagnostic performance compared with imaging. Results— Of a total of 565 patients, atypical presentations occurred in 31 LVO (38% of LVO) and 50 non-LVO cases (10%). Most scales correctly identified >95% of typical presentations but <20% of atypical presentations. Misclassification attributable to atypical presentations would have resulted in 4 M1/internal carotid artery occlusions, with National Institutes of Health Stroke Scale score ≥6 (5% of LVO) being missed and 9 non-LVO infarcts (5%) bypassing the nearest thrombolysis center. Conclusions— Atypical presentations accounted for the bulk of scale misclassifications, but the majority of these misclassifications were not detrimental, and use of LVO scales would significantly increase timely delivery to endovascular centers, with only a small proportion of non-LVO infarcts bypassing the nearest thrombolysis center. Our findings, however, would require paramedics to score as accurately as doctors, and this translation is made difficult by weaknesses in current scales that need to be addressed before widespread adoption.