To provide data guiding long-term antithrombotic therapy after coronary artery by-pass grafting (CABG) in patients with preoperative atrial fibrillation (AF).

From the SWEDEHEART registry, we included all patients, between January 2006 and September 2016, with preoperative AF and CHA₂DS₂-VASC score ≥2, undergoing CABG. Based on dispensed prescriptions 12 to 18 months after CABG, patients were divided in 3 groups: use of platelet inhibitors (PI) only, oral anticoagulant (OAC) only or a combination of OAC + PI. Outcomes were: Major adverse cardiac and cerebrovascular events (MACCE, [all-cause death, myocardial infarction, or stroke]), net adverse clinical events (NACE, [MACCE or bleeding]) and the individual components of NACE. Inverse probability of treatment weighting was used to adjust for the non-randomized study design.

Among 2,564 patients, 1,040 (41%) were treated with PI alone, 1,064 (41%) with OAC alone, and 460 (18%) with PI + OAC. Treatment with PI alone was associated with higher risk for MACCE (adjusted HR 1.43, 95% CI 1.09-1.88), driven by higher risk for stroke and MI, compared with OAC alone. Treatment with PI + OAC, was associated with higher risk for NACE (adjusted HR 1.40, 95% CI 1.06-1.85), driven by higher risk for bleeds, compared with OAC alone.

In this real-world observational study, a high proportion of patients with AF, undergoing CABG, did not receive a long-term OAC therapy. Treatment with OAC alone was associated with a net clinical benefit, compared with PI alone or PI + OAC.