Site Variation and Outcomes for Antithrombotic Therapy in Atrial Fibrillation Patients After Percutaneous Coronary Intervention

Background: Patients with atrial fibrillation (AF) treated with percutaneous coronary intervention (PCI) require multiple antithrombotic therapies. The optimal strategy is debated suggesting increased treatment variation. This study sought to characterize site-level variation in antithrombotic therapies in AF patients after PCI and determine the association with outcomes. Methods: Using the retrospective TREAT-AF study (The Retrospective Evaluation and Assessment of Therapies in AF) from the Veterans Health Administration, patients with newly diagnosed, nonvalvular AF between 2004 and 2015 followed by a PCI with a P2Y 12 -antagonist prescription were identified. Patients were grouped according to the therapy dispensed 7 days before until 30 days after the PCI: oral anticoagulation plus platelet inhibition (OAC+PI) or platelet inhibition only. A combined outcome of death, myocardial infarction, stroke, or major bleeding was assessed 1 year after PCI and Cox regression was performed to estimate hazard ratios. Results: Of 230 762 patients with newly diagnosed AF, 4042 (1.8%) underwent PCI and received a P2Y 12 -antagonist during the observation period (age, 67±9 years; CHA 2 DS 2 -VASc, 2.7±1.7; HAS-BLED, 2.6±1.2). Among these, 47% were prescribed OAC+PI, and 53% platelet inhibition only 7 days before until 30 days after the PCI. Across 63 sites, the use of OAC+PI ranged from 19% to 66%. Prescription of OAC+PI was independently associated with a reduction in the combined outcome of death, myocardial infarction, stroke, or major bleeding compared with platelet inhibition only (adjusted hazard ratio, 0.85; 95% CI, 0.73–0.99; P =0.033). Conclusions: In patients with established AF undergoing PCI, the use of OAC+PI varied substantially across sites in the 30 days post-PCI. Anticoagulation appeared to be underutilized but was associated with improved outcomes. Strategies to promote OAC+PI and minimize site variation may be useful, particularly in light of recent randomized trials.