Paroxysmal atrial fibrillation, stroke risk and thromboprophylaxis.

Thromb Haemost 2008; 100: 11–13 Atrial fibrillation (AF) is the commonest sustained cardiac rhythm disorder, which has particularly attracted much clinical and research interest from the mortality and morbidity associated with stroke and thromboembolism. The more chronic presentations of AF have clinical subtypes based on the temporal pattern of the arrhythmia – AF is considered recurrent when a patient experiences two or more episodes, which may be paroxysmal if they terminate spontaneously, defined by consensus as seven days, or persistent if the arrhythmia requires cardioversion for termination of the arrhythmia (1). Although paroxysmal AF is common, the associated stroke risk is less defined, given the limited information on its clinical epidemiology and the under-representation of these patients in clinical trials. As Gladstone et al. (2) highlight, it really depends on how hard we look for paroxysmal AF in patients who present with an acute stroke or thromboembolism. It has long been recognized that paroxysmal AF is frequently asymptomatic, and it has been estimated that only 1 in 12 paroxysms of AF are symptomatic (3). In their data from the prospective Registry of the Canadian Stroke Network, Gladstone et al. (2) found that 17% of 12,849 consecutive hospitalised ischemic stroke patients had a history of AF, and an additional 6% had new AF detected whilst in hospital (4% on admission electrocardiogram [ECG] and 2% detected later in hospital). In a study by Jabaudon et al. (4), a standard 12-lead ECG identified AF in 2.7% of 149 acute stroke and transient ischemic attack (TIA) patients at admission and in 4.1% of remaining patients within five days; however, a 24-hour ECG recording (Holter) detected AF in 5% of patients who had a normal standard ECG, whereas seven-day ambulatory ECG monitoring using an event-loop recording (ELR) device detected AF in 5.7% of patients with a normal standard ECG and normal Holter. Unsurprisingly, a recent systematic review of a series of cohort studies reemphasises the importance of screening for occult AF after ischemic stroke and TIA given that an extended duration of ECG monitoring (e.g. event loop recorder [ELR], etc) after stroke/TIA identifies many more patients with (usually paroxysmal) AF than standard ECG and Holter monitoring alone (5). Another recent report found that in patients with acute ischemic stroke, even frequent atrial premature beats (APBs) recorded in 24-hour ECG (≥70 per 24 hours) were a marker for subsequent paroxysmal AF (6). Clearly, more rhythm monitoring investigations would be needed for subjects with acute stroke and/or thromboembolism, to define the relationship between paroxysmal atrial arrhythmias and stroke.