Abstract TP114: Stroke Mimics: Uncovering The Truth In The History

Abstract

Background: Inpatient acute stroke identification may be challenging given the complexity related to both concomitant and comorbid illness. Rapid and accurate diagnosis requires a skilled clinician to discern acute stroke patients from other mimics such as hypoglycemia, hemiplegic migraine, post-ictal paresis, relapse of a demyelinating disease, or recrudescence of prior stroke. Hypothesis: Negative predictors of acute stroke can be identified in inpatients with inpatient stroke alerts. Methods: IRB approved, retrospective analysis of prospectively obtained data was performed for all inpatient stroke alerts between March 2017 and May 2020. Patients with MRI confirmed diagnosis of acute stroke were compared to all others. Univariant analysis was performed with Wilcoxon Ranked Sum test used for non-parametric data to identify risk factors to include in the logistic regression (p<0.05). Results: 674 inpatient stroke alerts resulted in 223 (33%) acute strokes and 451 (67%) non-stroke diagnoses. For stroke vs. non-stroke patients, median age 70.5 vs. 68.6 years, last seen well 77.5 vs. 65.0 minutes (p= 0.030), NIHSS score 10 vs. 4 (p<0.001). In the non-stroke group, 156 (23%) had a witnessed seizure documented during the alert compared to 0% in the stroke group. Negative predictors of stroke included previous diagnosis of epilepsy (0.302 odds, P<0.007), anti-seizure drug (ASDs) prescribed prior to (1.662 odds, p<0.001) or at time of the alert (2.479 odds), and previous EEG(P<0.001). Conclusions: It is imperative to consider post-ictal paresis as a differential diagnosis when assessing patients who present as stroke alerts. The administration of thrombolytics in stroke mimics is not without harm and incorrect diagnosis will delay appropriate treatment. Witnessed seizure at time of alert, history of epilepsy diagnosis, use of ASD prior to or at the time of the alert, and prior EEG findings may serve as a tool to help differentiate between acute stroke and seizure diagnosis. Further study will be aimed at identifying alternative pathways for rapid assessment of patients with acute changes in neurologic exam.