AFIRE - Antithrombotics for AF + stable CAD

in ,

Yasuda S, et al. Antithrombotic therapy for atrial fibrillation with stable coronary disease. NEJM 2019;381:1103-13.

Bottom Line:

  • Among Japanese patients with AF warranting OAC & stable CAD, using OAC monotherapy reduced the risk of dying (3% fewer) & having a major bleed (2% fewer) without increasing thrombotic risk compared with OAC plus an antiplatelet at 2 years.

  • These results are likely generalizable to non-Japanese patients & to other DOACs using approved doses.

Participants (n=2246 randomized, n=2215 analyzed)

  • Setting: Japan, enrolled 2015-2017

  • Included

    • Age >20 y

    • AF with CHADS2 score 1-6

    • Stable CAD (prior PCI or CABG >1 y ago, coronary stenosis 50%+ on angiography)

  • Key exclusion

    • Prior stent thrombosis

    • Active tumor

    • Poorly-controlled HTN

  • Baseline

    • Age 74 y, female 21%

    • AF: Paroxysmal 53%, persistent 15%, permanent 32%

    • CHADS2=2, CHA2DS2-VASc=4, HASBLED=2 (medians)

    • Previous stroke 15%, MI 35%, PCI 71% (drug-eluting stent 65-70%, mostly everolimus-eluting)

    • Current smoker 13%, diabetes 42%

    • CrCl <50 mL/min 34%

Intervention: Oral anticoagulant (OAC) monotherapy

  • Rivaroxaban 15 mg/d if CrCl 50+ (standard dose in Japan)

    • If CrCl 15-49: 10 mg/d (equivalent to 15 mg/d in non-Japanese)

    • 3.5% used antiplatelet

Control: OAC + antiplatelet

  • Rivaroxaban (as above) + antiplatelet (ASA or P2Y12 inhibitor at clinician’s discretion)

  • Antiplatelet used: ASA 70%, P2Y12 inhibitor 27%, other 2%, none <1%

Outcomes @ median 24 months

  • Primary outcome (death, stroke, systemic embolism, MI, unstable angina requiring PCI/CABG):

    • OAC 8.0% vs OAC+antiplatelet 10.9% (ARR 2.9%, NNT 35)

    • Hazard ratio (HR) 0.72 (95% confidence interval 0.55-0.95)

  • Death: 3.7% vs 6.6% (HR 0.55, 0.38-0.81; ARR 2.9%, NNT 35)

  • Ischemic stroke: 1.9% vs 2.5%

  • MI: 1.2% vs 0.7%

  • Major bleed (ISTH definition): 3.2% vs 5.2% (HR 0.59, 0.39-0.89; ARR 2%, NNT 50)

    • Hemorrhagic stroke: 0.4% vs 1.2% (HR 0.30, 0.10-0.92)

Internal Validity: Low risk of bias overall

  • Allocation bias: Low risk

    • Randomization using computer-generated minimization algorithm

    • Allocation concealed via web-based response system

  • Performance bias: Unclear risk

    • Open-label (patients & clinicians aware of allocated intervention)

    • Potential for crossover, differential management of antithrombotics & other cardiovascular risk factor modification

  • Detection bias: Low risk

    • Difference in outcomes driven by mortality; hard outcome with little opportunity to “game”

    • Outcomes assessed by blinded adjudication committee

  • Attrition bias: Low risk

    • Analyzed modified intention-to-treat (mITT) population & per-protocol (for non-inferiority)

    • Loss-to-follow-up: OAC monotherapy 2.5%, OAC+antiplatelet 1.8%

  • Other biases: Low risk

    • Stopped early for safety (higher risk of death in comparator group)

Other Considerations

  • Non-inferiority trial

    • Non-inferiority margin HR<1.46

    • Non-inferiority (and superiority) of OAC monotherapy met in both mITT & per-protocol analyses

  • Lower dose of rivaroxaban than used in North America

    • The approved “full dose” of rivaroxaban for stroke prophylaxis in AF in Japan is 15 mg/d (rather than 20 mg/d outside Asia) based on the J-ROCKET AF trial, as well as supporting pharmacokinetic-pharmacodynamic data

    • The 20 mg/d dose should be used among non-Asians. It is unclear what dose we should use in non-Japanese Asians & South Asians.