Embolic Strokes of Undetermined Source: An Outcome Analysis

Study Questions:

What are the outcomes of patients who have an embolic stroke of undetermined source (ESUS)?


This study used data from the Athens Stroke Registry between June 1992 and December 2011. This registry tracks patients with a first-ever ischemic stroke hospitalized at a single hospital in Athens. ESUS was defined as a visualized, on imaging, nonlacunar brain infarct in the absence of atherosclerotic disease causing ≥50% luminal stenosis, major-risk cardioembolic source, and any other specific cause of stroke. Other stroke types (e.g., cardioembolic, lacunar, etc.) were defined using standard definitions. All patients had a computed tomography (CT) on admission and CT or magnetic resonance imaging (MRI) 7-10 days later. Monitoring for arrhythmia was done with an electrocardiogram (ECG) on admission; continuous ECG monitoring for 7 days, or until discharge, for patients in the stroke unit; and 24 hours of ambulatory ECG monitoring in patients where atrial fibrillation (AF) was suspected. The primary outcome of the study was all-cause mortality, with secondary outcomes of stroke recurrence, favorable functional outcome (defined as an mRS ≤2), and a composite cardiovascular endpoint. Patients received follow-up at 1, 3, and 6 months after discharge and then annually. Patients who experienced more than one event during follow-up were censored after the first event. Outcomes by stroke type were estimated using the Kaplan-Meier limit method with differences evaluated with the log-rank test. Cox regression analyses were used to determine if stroke type predicted outcome and adjusted for various risk factors.


There were 2,731 patients included in the analysis and 275 (10%) had ESUS. The average follow-up was 30.5 ± 24.1 months, and 307 (11.2%) patients were lost to follow-up. Most ESUS patients (73.5%) were treated with an antiplatelet agent, although 16.7% received an anticoagulant, 5.3% were treated with a combination of an antiplatelet agent and an anticoagulant, and 4.5% received no antithrombotic therapy. The cumulative probability of survival was higher in the ESUS group: 65.6% (95% confidence interval [CI], 58.9%-72.2%) than the cardioembolic group: 38.8% (95% CI, 34.9%-42.7%) and the undetermined other than ESUS group: 46.4% (95% CI, 40.1%-52.7%). Cumulative survival was similar between the ESUS group and the large artery atherosclerotic group: 72.8% (95% CI, 68.3%-77.3%), and lower than the lacunar group: 81.0% (95% CI, 77.1%-84.9%). In the adjusted analyses, the risk of mortality was higher in patients with cardioembolic stroke (adjusted hazard ratios [aHR], 1.67; 95% CI, 1.29-2.15) and undetermined stroke other than ESUS (aHR, 1.87; 95% CI, 1.41-2.48) compared to patients with ESUS.

Stroke recurrence was more likely in patients with ESUS and cardioembolic strokes (cumulative probabilities, 29.0% [95% CI, 22.3%-35.7%] and 26.8% [95% CI, 22.1%-31.5%], respectively) than other stroke subtypes. In the adjusted analysis, the risk of stroke recurrence was lower in patients with large artery atherosclerotic strokes (aHR, 0.57; 95% CI, 0.41-0.83), lacunar strokes (aHR, 0.59; 95% CI, 0.43-0.83), and strokes of undetermined type other than ESUS (aHR, 0.55; 95% CI, 0.37-0.83) than in patients with ESUS.

Patients in the ESUS group had a favorable functional outcome (62.5%) in similar numbers to patients in the large artery atherosclerotic group (60.9%), but less than in the lacunar stroke group (82.2%). Patients with cardioembolic strokes were least likely to have a good functional outcome (32.2%).

Patients reached the composite cardiac endpoint more often in the ESUS and cardioembolic groups (cumulative probabilities, 38.1% [95% CI, 31.1%-45.2%] and 38.2% [95% CI, 33.3%-43.1%], respectively) than patients with other stroke subtypes (cumulative probabilities ranged from 24.3-29.8%). Patients with lacunar strokes were the only subtype significantly less likely to reach a composite cardiac event than patients with ESUS in the adjusted analysis (aHR, 0.70; 95% CI, 0.60-1.16).


Patients with ESUS have lower risk of mortality than patients with cardioembolic stroke, but a relatively high risk of recurrence, when compared with other stroke types.


Patients with ESUS present a management dilemma. They are at high risk of developing AF (>40% in some studies), but anticoagulation is not generally the first-line treatment, unless a clear indication is identified. This study suggests that patients with ESUS, while not at high risk of mortality compared with other stroke subtypes, are at an elevated risk of stroke recurrence. These findings raise the possibility that patients with ESUS should be empirically anticoagulated to reduce the risk of stroke recurrence, as a large proportion likely has paroxysmal AF. While this work is intriguing, it is a single-center retrospective study, and more work is needed in this area. Technology that allows detection of paroxysmal AF with devices such as implantable cardiac monitors was not available when this study took place. The paradigm of anticoagulating patients with ESUS, if AF is not identified, is currently being studied in randomized trials, and the results of that work should help to inform management of this condition.

Clinical Topics: Anticoagulation Management, Arrhythmias and Clinical EP, Noninvasive Imaging, Anticoagulation Management and Atrial Fibrillation, Implantable Devices, SCD/Ventricular Arrhythmias, Atrial Fibrillation/Supraventricular Arrhythmias, Magnetic Resonance Imaging

Keywords: Arrhythmias, Cardiac, Anticoagulants, Atherosclerosis, Atrial Fibrillation, Constriction, Pathologic, Electrocardiography, Fibrinolytic Agents, Magnetic Resonance Imaging, Platelet Aggregation Inhibitors, Recurrence, Risk Factors, Stroke, Stroke, Lacunar, Tomography

< Back to Listings