Background—Stroke is a debilitating condition with an increased risk in patients with atrial fibrillation. Although data from clinical trials suggest that both rate and rhythm control are acceptable approaches with comparable rates of mortality in the short term, it is unclear whether stroke rates differ between patients who filled prescriptions for rhythm or rate control therapy.

Methods and Results—We conducted a population-based observational study of Quebec patients ≥65 years with a diagnosis of atrial fibrillation during the period 1999 to 2007 with the use of linked administrative data from hospital discharge and prescription drug claims databases. We compared rates of stroke or transient ischemic attack (TIA) among patients using rhythm (class Ia, Ic, and III antiarrhythmics), versus rate control (β-blockers, calcium channel blockers, and digoxin) treatment strategies (either current or new users). The cohort consisted of 16 325 patients who filled a prescription for rhythm control therapy (with or without rate control therapy) and 41 193 patients who filled a prescription for rate control therapy, with a mean follow-up of 2.8 years (maximum 8.2 years). A lower proportion of patients on rhythm control therapy than on rate control therapy had a CHADS₂ (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, and previous stroke or TIA) score of ≥2 (58.1% versus 67.0%, P<0.001). Treatment with any antithrombotic drug was comparable in the 2 groups (76.8% in rhythm control versus 77.8% in rate control group). Crude stroke/TIA incidence rate was lower in patients treated with rhythm control in comparison with rate control therapy (1.74 versus 2.49, per 100 person-years, P<0.001). This association was more marked in patients in the moderate- and high-risk groups for stroke according to the CHADS₂ risk score. In multivariable Cox regression analysis, rhythm control therapy was associated with a lower risk of stroke/TIA in comparison with rate control therapy (adjusted hazard ratio, 0.80; 95% confidence interval, 0.74, 0.87). The lower stroke/TIA rate was confirmed in a propensity score–matched cohort.

Conclusions—In comparison with rate control therapy, the use of rhythm control therapy was associated with lower rates of stroke/TIA among patients with atrial fibrillation, in particular, among those with moderate and high risk of stroke.