Abstract Body: Introduction
Atrial fibrillation (AF), which can be easily identified through electrocardiogram (ECG), is a well-established predictor of ischemic stroke (IS) in comparison to other stroke subtypes. However, there is a paucity of literature surrounding other objective ECGs features and their discriminative abilities between stroke subtypes. This study explores the prevalence of ECG abnormalities and their potential associations with stroke and its subtypes.
Methods
Our analysis included 10,363 IS cases, 3,055 hemorrhagic stroke (HS) cases, and 13,488 matched controls from the INTERSTROKE case-control study. The ECG abnormalities investigated included AF, recent ST-elevation myocardial infarction (STEMI), left ventricular hypertrophy (LVH), right ventricular hypertrophy (RVH), P-pulmonale, and P-mitrale. Multivariate logistic regression, and multinomial regression by stroke severity, was performed and adjusted for covariates. Analyses were conducted using R version 4.3.1.
Results
In the population presenting with stroke, 19.26% exhibited ECG abnormalities (19.69% IS, 18.10% HS) compared to 7.1% of the controls. AF (7.4%) and LVH (11%) were the most reported abnormalities with recent STEMI, RVH, P-pulmonale and P-mitral each having a ≤1% prevalence (table 1). Regional variation was observed, with Africa having the highest detection of LVH (18%), and Western Europe/North America/Australasia having the highest detection of AF (9.2%).
Univariate and multivariate logistic regression analysis revealed AF as the strongest predictor of IS (OR 4.80, 95% CI 4.02–5.74), while LVH was the strongest predictor of HS (OR 5.35, 95% CI 4.06–7.04) (table 2). AF on presentation, significantly reduced the odds of a HS diagnosis (OR 0.22, 95% CI 0.14–0.33) compared to IS, while the converse was true of LVH (OR 1.76, 95% CI 1.35–2.29). Patients with LVH were significantly younger in age at stroke onset (61.0 ± 13.3 years vs. 62.3 ± 13.6 years, p < 0.001), particularly in the HS cohort (57.4 ± 12.8 years). LVH was most associated with severe stroke, defined as a modified Rankin score (mRS) of 4-6 on presentation (OR 1.20, 95% CI 1.03–1.41), compared to matched mild stroke cases.
Conclusion
ECG abnormalities are commonly detected in patients presenting with stroke, and substantive differences between stroke subtypes was demonstrated. These findings suggest that ECG could play a critical role in rapid stroke differentiation and aid in the development of future clinical prediction tools.