Atrial Fibrillation in Cryptogenic Stroke

Published - Written by Carla Rothaus

Your 69-year-old patient has just had an ischemic stroke. Fortunately, he should make a good recovery. And yet, you feel unsettled. A standard post-stroke work up – 12-lead ECG, ambulatory 24-hour Holter ECG monitoring, brain and neurovascular imaging, and echocardiography – failed to reveal a cause for his stroke, putting it in the “cryptogenic” stroke category. You worry though, that he might have intermittent episodes of atrial fibrillation you simply have been unable to detect….

Patients who have had a stroke or transient ischemic attack (TIA) caused by atrial fibrillation are at significant risk for recurrent stroke, yet atrial fibrillation is often asymptomatic and intermittent (paroxysmal), and may escape detection with short periods of ECG monitoring. Distinguishing strokes caused by atrial fibrillation from those without identifiable cause has important therapeutic implications. While anti-platelet therapy is the recommended treatment for cryptogenic stroke, it is much less effective than anticoagulation in preventing recurrent stroke when the cause is atrial fibrillation. Multiple observational studies have suggested that prolonged ECG monitoring is superior to conventional 24-hour ECG monitoring in identifying atrial fibrillation in patients with stroke. Data from randomized clinical trials have been lacking however, and new practice guidelines have not been adopted.

This week’s NEJM includes the results of two studies designed to address this void. Gladstone et al. conducted a multicenter trial (EMBRACE) that randomized 572 patients to undergo ambulatory monitoring with either a 30-day event-triggered external loop recorder or one additional round of 24-hour Holter monitoring. Eligible patients were 55 years or older, had no history of atrial fibrillation, and had been given a diagnosis of cryptogenic stroke or TIA after standard work up was negative. The primary outcome was detection of one or more episodes of ECG-documented atrial fibrillation or flutter lasting 30 seconds or longer within 90 days after randomization. A multicenter randomized controlled trial performed by Sanna et al. (CRYSTAL AF) compared prolonged monitoring with an implantable loop recorder to conventional ECG monitoring in 441 patients 40 years or older with recent cryptogenic stroke or TIA.

Both studies showed that long-term monitoring in the intervention group resulted in a higher rate of detection of atrial fibrillation. In EMBRACE, atrial fibrillation was detected in about 1 out of 6 patients in the intervention group, as compared to about 1 out of 30 in the control group; the difference was significant at the P<0.001 level. In CRYSTAL AF, the rate of detection of atrial fibrillation was about 1 out of 11 among patients in the intervention group, as compared to less than 1 in 50 among patients in the control group; this difference was also significant (P<0.001).

In an accompanying editorial, Dr. Hooman Kamel (Weill Cornell Medical College) calls the study findings an “important advance,” and concludes that “prolonged monitoring of heart rhythm should now become part of the standard care of patients with cryptogenic stroke.” He notes however that prior studies establishing the benefit of antithrombotic therapy for paroxysmal atrial fibrillation included patients whose arrhythmia was detected with conventional monitoring; additional research is needed, he cautions, to assure that brief episodes of paroxysmal atrial fibrillation detected with prolonged monitoring present a similar stroke risk and merit similar treatment.

NEJM Executive Editor Dr. Gregory Curfman adds, “These studies support the application of prolonged ECG monitoring for the detection of paroxysmal atrial fibrillation in patients with stroke. This approach may allow more targeted use of anticoagulation for stroke prevention.”