By Matthew E. Fink, MD

Professor and Chairman, Department of Neurology, Weill Cornell Medical College, Neurologist-in-Chief, New York Presbyterian Hospital

Dr. Fink reports no financial relationships relevant to this field of study.

Source: Favilla CG, et al. Predictors of finding occult atrial fibrillation after cryptogenic stroke. Stroke 2015;46:1210-1215.

In large prospective databases of ischemic stroke, cryptogenic stroke occurs in as many as 30% of all ischemic stroke patients. Many of these patients have the characteristics of cardioembolic stroke, and there has always been a strong suspicion that some of these patients have atrial fibrillation. However, only about 5% of patients with acute ischemic stroke are found to have atrial fibrillation while inpatients. Therefore, there is great interest in performing mobile cardiac outpatient telemetry (MCOT) after discharge, and this study looked at a retrospective cohort of consecutive patients who underwent 28-day MCOT after discharge.

There were 227 patients with cryptogenic stroke (179) or transient ischemic stroke (48), and 14% had atrial fibrillation detected on mobile cardiac outpatient telemetry. In a multivariate analysis of clinical, echocardiographic, and radiographic features associated with the strokes, the only significant independent predictors of finding paroxysmal atrial fibrillation were age > 60 years (odds ratio = 3.7) and prior cortical or cerebellar infarction seen on neuroimaging (odds ratio = 3.0). No other clinical features, including demographics, CHADS 2 score, congestive heart failure, hypertension, age, diabetes, prior stroke or TIA, vascular disease, sex, or stroke symptoms were significant predictors, nor did electrocardiographic findings or radiographic characteristics of the acute infarction have any significant association with the detection of paroxysmal AF. Therefore, these results would suggest monitoring patients over the age of 60 with evidence of prior stroke on brain imaging to look for atrial fibrillation. In this series, AF was detected in 33% of patients who had both of these features, but in only 4% of patients with neither.