SUBCLINICAL ATRIAL FIBRILLATION IS COMMON IN PATIENTS WITH CRYPTOGENIC STROKE

An implantable loop recorder established new diagnoses of atrial fibrillation in 25% of patients with cryptogenic stroke.

One third of ischemic strokes remain cryptogenic even after thorough inpatient evaluation. Numerous studies suggest that some of these patients may have paroxysmal atrial fibrillation (AF) that remains undiagnosed during their stroke hospitalization. Failing to detect these cases of AF may result in suboptimal antithrombotic therapy. However, what type and duration of cardiac monitoring should be used to rule out subclinical AF remain unclear.

To address this question, investigators placed implantable loop recorders (ILRs) in 51 patients who had received standard stroke evaluations, including vascular imaging, echocardiography (transthoracic in all patients and transesophageal in 30), and at least 24 hours of Holter monitoring without evidence of AF. The ILR software automatically detected AF episodes 2 minutes in duration. Two cardiologists independently reviewed any AF episodes detected by ILRs.

ILRs were implanted an average of 174 days after stroke. In 13 patients (25.5%), AF was detected after a median 48 (range, 0–154) days of monitoring. The other patients remained AF-free throughout an average 229 days of monitoring. AF was associated with older age, more-frequent premature atrial contractions during baseline Holter monitoring, and larger left atrial size.

Comment: In the recently completed EMBRACE trial, 15% of cryptogenic stroke patients assigned to 30-day external loop recorder monitoring received new diagnoses of atrial fibrillation, versus 3% of patients assigned to 24-hour outpatient Holter monitoring. It is possible but unlikely that some of these poststroke episodes of AF are incidental, because subclinical AF lasting only a few minutes has recently been shown to increase stroke risk, and the majority of cryptogenic strokes appear radiographically to have resulted from cardiac embolism. On these bases, several weeks of noninvasive cardiac monitoring should usually be performed in patients with cryptogenic stroke. The current study results suggest that longer periods of monitoring might detect even more cases of AF. However, relatively invasive and expensive implantable loop recorders cannot be routinely recommended until we see the results of the ongoing CRYSTAL-AF trial, which is comparing use of ILRs to routine clinical follow-up in patients with cryptogenic stroke.