Accuracy of Stroke Diagnosis in Telestroke-Guided Tissue Plasminogen Activator Patients

Abstract

The objective of the study is to assess the accuracy of final diagnosis in telestroke-guided tissue plasminogen activator (rt-PA) patients compared with bedside evaluation using computed tomography (CT) or magnetic resonance imaging (MRI) as a surrogate for final stroke diagnosis. The overall goal was to determine if telestroke had similar diagnostic accuracy as bedside evaluations in diagnosing rt-PA-treated patients. We analyzed all acute stroke code calls who received intravenous rt-PA at our center from October 2013 to June 2015. Calls were grouped into patients who were initially evaluated by telestroke at a spoke partner site (spoke) and patients evaluated in person at the hub. Patients receiving additional neurointervention were excluded to avoid confounding. Relevant variables included severity of stroke (National Institutes of Health Stroke Scale [NIHSS]), rt-PA times, presence of intracranial hemorrhage (ICH), and primary outcome of CT or MRI evidence of stroke after rt-PA administration. Post-rt-PA imaging used included CT or MRI within 72 hours after treatment. Overall, 80 patients received intravenous rt-PA (spoke [n = 23] and hub [n = 57]). There was no difference in mean NIHSS score prior to treatment (10.3 ± 9.2 and 9.8 ± 8.4; P = .936), "onset-to-treatment" time (143.6 ± 53.5 minutes and 141.0 ± 54.1 minutes; P = .915), and ICH rate (13% and 8.8%; adjusted P = .898). The presence of radiographic evidence of stroke at spoke versus hub was not different (78.3% and 66.7%; adjusted P = .338). The most commonly used radiographic modality was MRI (MRI: 80%, CT: 20%). Using imaging as a surrogate for final diagnosis resulted in no difference in final stroke diagnosis rate between the groups, reinforcing that telestroke evaluations are as accurate as bedside evaluations in diagnosing acute stroke.