Abstract Body: Background: During the course of an ischemic stroke hospitalization, the initial diagnosis of the causative mechanism or etiology of stroke may evolve. We aimed to quantify the degree to which patients with initial etiologic diagnosis are reclassified during the hospitalization and frequencies of reclassified diagnoses.
Methods: This retrospective cohort consisted of all patients from a single center recorded in the institutional Get-with-the-guidelines Stroke registry from September 2015-December 2020 with a stroke etiology entered who were evaluated by at least 1 vascular neurologist. The stroke etiologies documented were mapped to one of the 5 TOAST classifications: 1-large artery atherosclerosis, 2-cardioembolism, 3-small vessel disease, 4-other determined etiology, 5-undetermined etiology. Spline models and a Sankey diagram were employed to visualize the trends in etiology diagnosis reclassification by subsequent vascular neurologists during the hospitalization.
Results: A total of 709 patients were included in the analytic cohort. Initial TOAST classifications were the following: TOAST 1 (n= 132, 19%), TOAST 2 (n= 168, 24%), TOAST 3 (n=88, 12%), TOAST 4 (n=50, 7%), TOAST 5 (n= 271, 38%). There were 132 (19%) patients who were reclassified with a different etiologic diagnosis by the time of hospital discharge. The number of patients who were reclassified from their initial etiologic diagnosis were: TOAST 1 (n=11), 2 (n=9), 3 (n=4), 4 (n=4), and 5 (n=104). The spline plots in the Fig. 1 demonstrate that an increasing number of impressions by subsequent vascular neurologists was associated with increasing likelihood of reclassification for subtypes 4 and 5. For each subtype, the most common reclassifications were TOAST 1 to 5 (54.5%), 2 to 1 (33.3%), 3 to 5 (50.0%), 4 to 5 (50.0%), and 5 to 2 (58.7%). Distributions of initial and final etiologic diagnoses among patients who had diagnostic reclassification are displayed in the Sankey diagram in Fig. 2.
Discussion: Identifying stroke etiology is important for ensuring that patients receive targeted secondary stroke prevention. Strokes with initial subtypes of 4 and 5 were more likely to receive a different diagnosis with an increasing number of impressions by subsequent vascular neurologists. The treatment profiles of the initial and reclassified subtypes are distinct, highlighting the value of inpatient diagnostic evaluation, particularly cardiac testing, during an acute stroke hospitalization.