Variables That Best Differentiate Acute Stroke from Stroke Mimics in Inpatients with Acute Neurological Deficits (P07.015)

Abstract

Objective: To identify inpatient variables that best differentiate acute strokes from stroke-mimics Background Identifying and triaging in-hospital strokes can be difficult but do represent a valuable treatment opportunity. We sought to systematically analyze symptoms, signs, and other characteristics of inpatient strokes with the hypothesis that focal neurological symptoms/deficits would best predict true strokes whereas altered mental status predict stroke-mimics. Design/Methods: We retrospectively analyzed prospectively collected inpatient stroke calls over one year (n=93) and retrospectively analyzed systematically collected data focusing on presenting symptoms, signs, stroke risk factors, and nursing unit location. We excluded patients admitted to a neurology service. Chi-square (categorical variables), t-test (continuous) and binary multiple logistic regression (multivariate analysis) were used to compare strokes versus stroke-mimics. Results: Fifty-seven (mean age 69.6 years; 54% female) of the 93 patients had a confirmed new stroke with 42.1% having an NIHSS 2 of model = 0.322]. 51.7% of the true strokes had atrial fibrillation at the time of stroke and 6.9% within 72 hours. Total NIHSS and laboratory values did not predict true stroke. Conclusions: The majority of acute inpatient stroke calls are for patients with true stroke. Most true stroke patients had atrial fibrillation, a complaint or finding of arm weakness, an NIH1b score of 0, and normal mental status. Disclosure: Dr. Newey has nothing to disclose. Dr. Wisco has nothing to disclose. Dr. George has nothing to disclose. Dr. Gebel has nothing to disclose.