Aim:
To study echocardiographic characteristics assessed by transthoracic and transoesophageal echocardiography in patients with acute ischemic stroke or transient ischemic attack (TIA) and their potential association with stroke aetiology according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria.

Methods:
We analysed echocardiographic data from the prospective, multicentre CONTEST-Study (Comparative Effectiveness Study of Transthoracic and Transoesophageal Echocardiography in Stroke). Over a two-year period, 494 ischemic stroke or TIA patients with undetermined aetiology at the time of enrolment were included at five certified stroke centres in Germany. After exclusion of individuals with missing or insufficient information on echocardiography or TOAST-classification 475 individuals were included in the current analysis. Furthermore, we excluded individuals with suspected paradoxical embolism due a patent foramen ovale (N=25) and dichotomized stroke subgroups into presumed cardioembolic versus non-cardioembolic causes. Echocardiographic assessment included various parameters of left ventricular systolic and diastolic function as well as atrial volumes. We performed multivariable adjusted logistic regression analyses in order to determine a potential predictive ability of specific echocardiographic parameters for cardioembolic stroke.

Results:
The median age of the individuals was 65.0 years (interquartile range 22 years), 58.9% were men. 423 (94.0%) individuals were classified as non-cardioembolic and 27 (6%) as cardioembolic ischemic stroke/TIA. Patients with a cardioembolic stroke/TIA showed a larger left atrial volume and a higher mitral early inflow peak velocity compared to individuals with non-cardioembolic. In fully adjusted regression analyses, increasing left ventricular ejection fraction was the only echocardiographic parameter associated with a decreased risk of cardioembolic stroke/TIA (odds ratio [OR] 0.95, 95% confidence interval [CI] 0.90-1.00, p=0.046). Left atrial volume index and the E/A-ratio showed a trend towards an increasing risk for cardioembolic stroke/TIA with higher values, but did not meet statistical significance (OR 3.62 [95% CI 0.79-16.6; p=0.098] and 1.99 [95% CI 0.87-4.55; p=0.104], respectively).

Conclusion:
Individuals with cardioembolic stroke/TIA had larger left atria and a higher mitral early inflow peak velocity compared to individuals with non-cardioembolic stroke/TIA. However, after adjustment for cardiovascular comorbidities and risk factors, only left ventricular ejection fraction remained inversely associated with cardioembolic acute ischemic stroke/TIA.