Diagnostic Yield of Echocardiography in Transient Ischemic Attack

Abstract

Echocardiography is often performed to identify a cardiac source of embolism (CSE) causing transient ischemic attack (TIA). However, the diagnostic yield of echocardiography in TIA remains uncertain, and its role in routine evaluation of TIA is controversial. Patients with acute TIA were prospectively enrolled at 4 stroke centers. A CSE was defined using the Causative Classification of Stroke system; patent foramen ovale was considered a relevant CSE only if the patient underwent closure or was placed on anticoagulation. Patients with a known CSE at time of admission were excluded from analysis of the yield of echocardiography. A total of 869 patients were enrolled at stroke centers, and 129 had a known CSE at presentation. Of the 740 remaining patients, 603 (81%) underwent echocardiography. A potential CSE was identified in 60 (10%) of these patients. The most common CSEs noted on echocardiography were complex aortic arch atherosclerosis and patent foramen ovale. History of coronary artery disease (P < .001), lack of prior stroke or TIA (P = .007), and presence of acute infarction on magnetic resonance imaging (MRI) (P < .001) were predictors of CSE on echocardiography. The yield of echocardiography was 29% in patients with both history of coronary artery disease and acute infarction on MRI, 14% with one of these features, and 5% with neither of these features (P < .0001). A CSE identified by echocardiography prompted initiation of anticoagulation in 15 of the 603 (2.5%) subjects. Echocardiography demonstrates a relevant CSE in a significant portion of patients with TIA. However, changes in antithrombotic therapy resulting from echocardiography are infrequent.