The purpose of this study is to compare the efficacy and safety of rivaroxaban with warfarin for the prevention of blood clots in the brain (referred to as stroke) and blood clots in other parts of the body referred to as non-central nervous system systemic embolism) in patients with non-valvular atrial fibrillation (a heart rhythm disorder).

Patients with non-valvular atrial fibrillation who are at risk for stroke and non-central nervous system (non-CNS) systemic embolism, will be randomized (assigned by chance) to receive treatment with rivaroxaban or warfarin, two different anticoagulants (substances that prevent blood clots). Treatment will be double-blinded (neither the patient nor study staff will know which study drug is assigned to patients during the study). Patients assigned to rivaroxaban will receive rivaroxaban 20 mg orally (p.o.) once daily (OD) plus warfarin placebo p.o. OD titrated to a target sham international normalized ratio (INR) of 2.5. Patients with moderate renal impairment at screening will receive rivaroxaban 15 mg p.o. OD. Patients assigned to warfarin will receive warfarin p.o. OD titrated to a target INR of 2.5 plus rivaroxaban placebo p.o. OD. The maximum expected length of treatment is up to 32 months but may be extended up to 4 years.

Inclusion Criteria:

• Patients must have documented atrial fibrillation on 2 separate occasions within 6 months before screening
• History of a prior stroke, transient ischemic attack or non-neurologic systemic embolism believed to be cardiac in origin, or at least two of the following risk factors: heart failure, hypertension, age 75 years or greater, diabetes mellitus

Exclusion Criteria:

• Significant mitral stenosis
• Transient atrial fibrillation caused by a reversible disorder
• Active internal bleeding
• Severe disabling stroke
• History of intracranial bleeding
• Hemorrhagic disorders