Study design

Design:

individual patient data meta-analysis of 12 randomised controlled trials (RCTs). 6 RCTs were placebo-controlled; 8 RCTs compared an oral anticoagulant (OAC) with an antiplatelet (AP).

Blinding:

outcome adjudication committee was blinded in 11 trials.

Study question

Patients:

8932 adults (mean age 72 y, 63% men) with non-valvular atrial fibrillation. Patients with clinical indications for or against any of the active therapies were excluded.

Interventions:

full-dose OACs (mainly warfarin sodium or 4-hydroxycoumarin) with lower target international normalised ratio (INR) of 1.5–2.8 and upper target INR of 2.7–4.2 (n = 3430); APs (mainly acetylsalicylic acid, 75–325 mg) with or without low-dose OACs (median INR <1.5; n = 3531); or placebo (n = 1971).

Outcomes:

ischaemic stroke, serious bleeding, and cardiovascular events (ischaemic stroke, myocardial infarction, systemic embolism, or vascular death). Intention-to-treat analysis.

Follow-up period:

mean 2 years.

Main results

Older age was associated with increased risk of …