Risk of Ischemic Events in Persistent vs. Paroxysmal AF

Study Questions:

In patients with a history of ischemic stroke and atrial fibrillation (AF), is there a difference in the risk of future stroke between those with paroxysmal versus sustained atrial fibrillation?


This study used data from SAMURAI-NVAF, which was a multicenter, prospective, observational study that evaluated anticoagulant choices for patients with AF after ischemic stroke and transient ischemic attack (TIA). The current study includes patients who were hospitalized within 7 days of stroke/TIA between April 2011 and March 2014 at 18 Japanese stroke centers. AF was diagnosed by 12-lead electrocardiogram, cardiac telemetry, or the medical record. The attending physician adjudicated if the patient had paroxysmal AF or sustained (persistent and permanent) AF. Clinical and demographic information was abstracted. The primary endpoint was stroke or systemic embolism ascertained at 3 months, 1 year, and 2 years after the index event. Bleeding events and functional status were also ascertained.


There were 1,192 patients included in this analysis, and the median follow-up was 1.8 years. The average age was 77.7 ± 9.9 years, 44% were women, and 63.6% had sustained AF. More than 94% of patients were taking an anticoagulant, and the average CHADS2 score was 4. When compared to patients with paroxysmal AF, those with sustained AF were more likely to have congestive heart failure, prior stroke/TIA, history of hemorrhage, liver problems, higher alcohol use, and more disability after the index stroke. After adjustment for prognostic factors, paroxysmal AF was associated with increased odds of functional independence 3 months after the index stroke (adjusted odds ratio, 1.67; 95% confidence interval [CI], 1.18-2.37). Patients with sustained AF were more likely than those with paroxysmal AF to take warfarin, rather than a new oral anticoagulant. Stroke or systemic embolism was more common in patients with sustained AF than paroxysmal AF (8.3 per 100 person-years vs. 4.6 per 100 person-years). Bleeding was also more common in the sustained AF group (3.4 per 100 person-years vs. 3.1 per 100 person-years). After adjustment, sustained AF was an independent risk factor for stroke or systemic embolism (adjusted hazard ratio, 1.95; 95% CI, 1.26-3.14), but there was no increased bleeding risk with sustained AF in the adjusted analysis.


In patients with a history of ischemic stroke and AF, patients with sustained AF were more likely to have residual disability and recurrent stroke than patients with paroxysmal AF.


AF is an important stroke risk factor. Prior studies have had conflicting findings regarding the risk of recurrent embolic events in patients with sustained versus paroxysmal AF. The mechanism by which sustained AF is associated with a higher risk of stroke than paroxysmal AF is unclear. The primary limitation of this work is its susceptibility to selection bias, as it is an observational study. The finding that sustained AF is associated with a higher risk of stroke than paroxysmal AF in patients with a history of ischemic stroke is intriguing, and if confirmed by additional studies, could help stratify risk of stroke in patients with AF.

Keywords: Anticoagulants, Arrhythmias, Cardiac, Atrial Fibrillation, Electrocardiography, Embolism, Heart Failure, Hemorrhage, Ischemic Attack, Transient, Risk Factors, Secondary Prevention, Stroke, Telemetry, Vascular Diseases, Warfarin

< Back to Listings