In the article, “Optimizing in-hospital triage for large vessel occlusion using a novel clinical scale (GAI2AA),” Drs. Ohta et al. retrospectively analyzed a cohort of 429 patients with acute ischemic stroke from a single center and developed a clinical scale to predict a proximal anterior circulation large vessel occlusion (LVO), GAI2AA, which they then applied to a prospective cohort of 259 patients from 3 stroke centers for external validation. They also examined the utility of the scale for in-hospital triage and workflow factors for 158 patients with such LVOs. They found that the scale's performance was either better than or similar to other available scales, with a C-statistic of 0.90 and that the use of the scale was associated with shorter door-to-puncture times. In response, Dr. Heldner appreciates the novel features of the scale including the combined scoring of gaze palsy, aphasia, and inattention as well as the inclusion of atrial fibrillation (AF). Having retrospectively applied the scale to patients in the Bernese ischemic stroke database presenting within 6 hours of onset and using the recommended cutoff, she reports a C-statistic of 0.80. Discussing the practical implications and limitations that this would have in prehospital/interhospital transfer decisions for patients with suspected LVO, she notes that the true utility of such scores in this triage is debatable because some LVO cases would be missed. Responding to these comments, the authors note that a stroke-trained hospital physician scored the patients in their study on hospital arrival and that prehospital patient triage was beyond the scope of the study. They suggest that the GAI2AA scale should be used as in-hospital triage to facilitate rapid LVO-related workflow, rather than to achieve accurate prediction of LVO in prehospital settings. In another response, Drs. Grewal and Goldstein noted that using a different case definition of LVO (that included the posterior circulation) in patients admitted to the University of Kentucky Medical Center, they too found that a history of AF was associated with LVO, although less frequently than reported by the authors, and with no more than moderate discriminative capacity for LVO. Although this added item did not improve the accuracy of common prehospital scales in identifying LVO in their work, they acknowledge that identifying a history of AF may alert emergency responders of a potential LVO in a patient with suspected stroke. Responding to these comments, the authors note that AF had the weakest correlation with LVO and that including AF in the GAI2AA scale was intended to reinforce the probability of LVO in addition to suggesting a cardioembolic origin. This exchange reflects the ongoing discussion in the stroke literature about potential applications of LVO-related scores in stroke care and the challenges of constructing such scores in the first place.
In the article, “Optimizing in-hospital triage for large vessel occlusion using a novel clinical scale (GAI2AA),” Drs. Ohta et al. retrospectively analyzed a cohort of 429 patients with acute ischemic stroke from a single center and developed a clinical scale to predict a proximal anterior circulation large vessel occlusion (LVO), GAI2AA, which they then applied to a prospective cohort of 259 patients from 3 stroke centers for external validation. They also examined the utility of the scale for in-hospital triage and workflow factors for 158 patients with such LVOs. They found that the scale's performance was either better than or similar to other available scales, with a C-statistic of 0.90 and that the use of the scale was associated with shorter door-to-puncture times. In response, Dr. Heldner appreciates the novel features of the scale including the combined scoring of gaze palsy, aphasia, and inattention as well as the inclusion of atrial fibrillation (AF). Having retrospectively applied the scale to patients in the Bernese ischemic stroke database presenting within 6 hours of onset and using the recommended cutoff, she reports a C-statistic of 0.80. Discussing the practical implications and limitations that this would have in prehospital/interhospital transfer decisions for patients with suspected LVO, she notes that the true utility of such scores in this triage is debatable because some LVO cases would be missed. Responding to these comments, the authors note that a stroke-trained hospital physician scored the patients in their study on hospital arrival and that prehospital patient triage was beyond the scope of the study. They suggest that the GAI2AA scale should be used as in-hospital triage to facilitate rapid LVO-related workflow, rather than to achieve accurate prediction of LVO in prehospital settings. In another response, Drs. Grewal and Goldstein noted that using a different case definition of LVO (that included the posterior circulation) in patients admitted to the University of Kentucky Medical Center, they too found that a history of AF was associated with LVO, although less frequently than reported by the authors, and with no more than moderate discriminative capacity for LVO. Although this added item did not improve the accuracy of common prehospital scales in identifying LVO in their work, they acknowledge that identifying a history of AF may alert emergency responders of a potential LVO in a patient with suspected stroke. Responding to these comments, the authors note that AF had the weakest correlation with LVO and that including AF in the GAI2AA scale was intended to reinforce the probability of LVO in addition to suggesting a cardioembolic origin. This exchange reflects the ongoing discussion in the stroke literature about potential applications of LVO-related scores in stroke care and the challenges of constructing such scores in the first place.