ABSTRACT NUMBER: ESOC2026A528 REAL-WORLD PATIENT CHARACTERISTICS AND CLINICAL OUTCOMES ACROSS STROKE TOAST SUBTYPES: ANALYSIS OF A LARGE REPRESENTATIVE US ELECTRONIC HEALTH RECORD DATABASE

Abstract Background and aims Stroke subtypes per TOAST classification are predictors of prognosis and treatment response. However, limited real-world data exist across subtypes. We assessed real-world patient characteristics, treatment patterns and outcomes per TOAST subtypes. Methods Adults hospitalised with a first NCIS between 01/2017–12/2023 were included in this retrospective observational study using EHRs from the large, nationally representative US Truveta database. Patients with baseline diagnoses suggesting cardioembolic aetiology were excluded. Baseline patient characteristics, treatment patterns and 1-year outcomes were assessed by TOAST subtype (derived from unstructured clinical notes using Truveta’s LLM). Results Of 180,021 patients with NCIS, 17% (n=29,816) had a documented TOAST subtype. SVO was most common (30%), then stroke of undetermined aetiology (23%), LAA (22%), cardioembolic (16%), undetermined–combination (7%) and stroke of other determined aetiology (3%). Hypertension and hyperlipidaemia were the most common comorbidities across all subtypes (Figure). Median NIHSS scores ranged from 2–4. LAA strokes were most frequently treated with DAPT during admission (52.9% vs all others <42%) and at discharge (59.4% vs all others ≤48%). The incidence rate for recurrent IS at 1 year ranged from 12.5 to 15.67 for all subtypes except SVO (9.91). Stroke of other determined aetiology and cardioembolic strokes had the highest rates of major bleeding and ICH. Conclusions These real-world data indicate that clinical outcome incidence rates were consistent across all TOAST subtypes, except SVO, despite a numerically higher proportion with atherosclerotic disease risk factors, hypertension, diabetes and hyperlipidaemia. Conflict of interest This study was funded by Bayer AG. AS reports research grants/contracts from AstraZeneca, Bayer, the Canadian Institutes of Health Research, Daiichi Sankyo Company, the Heart and Stroke Foundation of Canada, the National Institutes of Health, Octapharma USA, Inc. and Servier Affaires Medicales; has served on advisory boards for AstraZeneca, Bayer, Daiichi Sankyo Company and Takeda Pharmaceutical Company; and has served on a data and safety monitoring board for Bayer. TNN reports serving as an Associate Editor of Stroke and consulting relationships with Bayer, Medtronic and Route92. MR declares no conflict of interest. SS’s institution received compensation from Bayer. MD and NT are employees of Truveta. SK, DRM, JX and KA are employees of Bayer. Figure 1 - belongs to Conclusions