The trial was designed and led by an academic steering committee whose members were responsible for the conduct of the trial. The Duke Clinical Research Institute (DCRI, Durham, NC) was the academic coordinating center. The trial was sponsored by Bristol-Myers Squibb and Pfizer. The trial protocol (available with the full text of this article at NEJM.org) and all subsequent amendments were approved by national regulatory agencies in participating countries and by institutional review boards or ethics committees at participating sites. An independent data and safety monitoring board reviewed unblinded patient-level data at regular intervals during the trial. Although Bristol-Myers Squibb assisted with data management, all the statistical analyses were performed independently at the DCRI. The initial draft of the manuscript was written by the first author and revised on the basis of comments from the other authors. All the authors vouch for the adherence of the trial to the protocol, and the first, second, and last authors vouch for the accuracy and completeness of the data and analysis. No one who is not an author contributed to writing the manuscript. The committee members and participating investigators are listed in the Supplementary Appendix, available at NEJM.org.
The trial rationale and design have been published previously.11 In brief, AUGUSTUS was a prospective, multicenter, two-by-two factorial, randomized clinical trial comparing apixaban with a vitamin K antagonist and comparing aspirin with placebo in patients with atrial fibrillation who had a recent acute coronary syndrome or underwent PCI (or both).
Patients who met all the following criteria were eligible for inclusion: an age of at least 18 years; previous, persistent, permanent, or paroxysmal atrial fibrillation and planned long-term use of an oral anticoagulant; recent acute coronary syndrome or PCI; and planned use of a P2Y12 inhibitor for at least 6 months. Patients who were using anticoagulation for other conditions (e.g., prosthetic valves, venous thromboembolism, and mitral stenosis) were not eligible. Other key exclusion criteria were severe renal insufficiency, a history of intracranial hemorrhage, recent or planned coronary-artery bypass graft surgery, coagulopathy or ongoing bleeding, and contraindication to a vitamin K antagonist, apixaban, all P2Y12 inhibitors, or aspirin. All the patients provided written informed consent before enrollment.
Stroke and bleeding risks were assessed with the use of two scores (Table S1 in the Supplementary Appendix). CHA2DS2-VASc scores reflect the risk of stroke among patients with atrial fibrillation who are not receiving anticoagulant therapy; scores range from 0 to 9, with higher scores indicating greater risk. HAS-BLED scores reflect the risk of bleeding among patients with atrial fibrillation who are receiving anticoagulant therapy; scores range from 0 to 9, with higher scores indicating greater risk.
Patients underwent randomization within 14 days after having an acute coronary syndrome or undergoing PCI, with explicit guidance provided to enroll eligible patients and start the trial regimen as soon as possible after the index event once parenteral anticoagulation had been stopped. To be eligible, patients were required to be planning to use an approved P2Y12 inhibitor for at least 6 months. The choice of P2Y12 inhibitor was left to the discretion of the treating physician.
After enrollment into this trial that had a two-by-two factorial design, patients were randomly assigned by means of an interactive voice-response system to receive apixaban or a vitamin K antagonist and to receive aspirin or matching placebo. The treatment regimen comparing apixaban with a vitamin K antagonist was open-label; however, the regimen comparing aspirin with matching placebo was double-blind.
Randomization was stratified according to indication (acute coronary syndrome or PCI) at enrollment. In accordance with the apixaban label instructions for stroke prevention in patients with atrial fibrillation, patients who had been randomly assigned to receive apixaban were directed to take 5 mg twice daily or to take 2.5 mg twice daily if they met two or more of the following dose-reduction criteria: were at least 80 years of age, had a weight of no more than 60 kg, or had a creatinine level of at least 1.5 mg per deciliter (133 μmol per liter). Patients who had been randomly assigned to receive a vitamin K antagonist had the dose adjusted to reach a target international normalized ratio (INR) within a range of 2.0 to 3.0. Regarding the comparison of aspirin with placebo, patients received aspirin at a dose of 81 mg or a matching placebo once daily. After 6 months, patients were transitioned from their two trial interventions to receive antiplatelet and anticoagulant therapy according to the local standard of care.
All the patients who underwent randomization were to be followed through 6 months, with an additional visit at month 7 to record transitions in antithrombotic therapy and associated outcomes. The primary outcome for both factorial comparisons was major or clinically relevant nonmajor bleeding as defined by the International Society on Thrombosis and Haemostasis (ISTH). ISTH major bleeding was defined as bleeding that resulted in death, occurred in a critical organ (intracranial, intraspinal, intraocular, retroperitoneal, intraarticular, intramuscular with compartment syndrome, or pericardial), or was associated with either a decrease in the hemoglobin level of at least 2 g per deciliter or a transfusion of at least 2 units of packed red cells.12 Clinically relevant nonmajor bleeding was defined as bleeding that resulted in hospitalization, medical or surgical intervention for bleeding, an unscheduled clinic visit, or a change in physician-directed antithrombotic therapy. Bleeding was also classified according to the Global Use of Strategies to Open Occluded Arteries (GUSTO) and Thrombolysis in Myocardial Infarction (TIMI) definitions (see the Supplementary Appendix).13,14
Secondary outcomes included the composite of death or hospitalization and the composite of death or ischemic events (stroke, myocardial infarction, stent thrombosis [definite or probable], or urgent revascularization). Exploratory outcomes included individual components of the secondary outcomes. All bleeding and ischemic events (except for urgent revascularization) were independently adjudicated by the clinical-events classification committee at the DCRI, whose members were unaware of the trial-group assignments.
The trial used a factorial design to evaluate two independent hypotheses. One hypothesis was that apixaban would be at least noninferior, and possibly superior, to a vitamin K antagonist with regard to the outcome of ISTH major or clinically relevant nonmajor bleeding in patients with atrial fibrillation and recent acute coronary syndrome or PCI and planned concomitant antiplatelet therapy. The second hypothesis was that single antiplatelet therapy with a P2Y12 inhibitor would be superior to dual antiplatelet therapy with a P2Y12 inhibitor and aspirin with regard to the outcome of ISTH major or clinically relevant nonmajor bleeding in patients with atrial fibrillation and recent acute coronary syndrome or PCI and planned concomitant anticoagulant therapy.
The population for the primary outcome analysis included all the patients who underwent randomization and received at least one dose of a trial drug or placebo. Events were counted from the beginning of receipt of the trial intervention through 2 days after the permanent discontinuation of the relevant drug or placebo. The population for the two secondary outcome analyses included all the patients who underwent randomization according to the randomized groups, and all events were counted from randomization through the 6-month visit.
Using a log-rank test to compare apixaban with a vitamin K antagonist, we estimated that 357 patients with primary-outcome events among 4600 patients who had undergone randomization would provide the trial with 77% power to detect a prespecified noninferiority relative margin of 1.2 at the upper one-sided 97.5% level of confidence, assuming a 1% annual loss to follow-up. Interaction between the two randomization factors (oral anticoagulant and aspirin) on the primary and secondary outcomes was assessed, with stratification according to the index event. If there was no significant interaction, the two factors were analyzed independently.
If noninferiority of apixaban as compared with a vitamin K antagonist was met, we used a prespecified hierarchical testing procedure by sequentially testing for the superiority of apixaban over a vitamin K antagonist for the following outcomes (in order): the composite of ISTH major or clinically relevant nonmajor bleeding (primary safety outcome), the composite of death or hospitalization, and the composite of death or ischemic events. If superiority was not met in any test, P values would not be reported for this or subsequent outcomes. P values were calculated for interaction tests of prespecified subgroups. Hierarchical testing was performed in the same way for aspirin as compared with placebo, except that the initial noninferiority test on the primary outcome was omitted. All P values for superiority were two-sided.
The analysis of the time to the first primary safety outcome was performed with the use of a Cox proportional-hazards model that included anticoagulant-regimen group (apixaban or vitamin K antagonist) as a covariate, stratified according to the index event (any acute coronary syndrome or elective PCI) and the antiplatelet-regimen group (aspirin or placebo). A similar comparison was performed for the time to the first primary safety outcome for the comparison of aspirin with placebo. Denominators for presented percentages represent nonmissing values, unless otherwise indicated. All the statistical analyses were performed with the use of SAS software, version 9.4 (SAS Institute).