Antithrombotic Strategy in AF with Stable Coronary Artery Disease
Definition
Patients with atrial fibrillation and concomitant stable coronary artery disease (stable CAD) present an antithrombotic dilemma: AF mandates oral anticoagulation to prevent stroke/systemic embolism, while CAD conventionally calls for antiplatelet therapy to prevent ischemic events. Combined OAC + antiplatelet (dual antithrombotic therapy) increases bleeding risk substantially. The clinical question is whether OAC monotherapy is sufficient — preserving ischemic protection while eliminating excess bleeding — in the chronic, stabilized phase (at least 6–12 months after the index coronary event or PCI).
This concept applies specifically to the stable/chronic CAD phase. For the early post-PCI/ACS period, see concepts/DAPT-Strategies and entities/Atrial-Fibrillation (ACS section).
Key Concepts
The Clinical Dilemma
- AF increases stroke risk 2–5×; long-term OAC is Class I. (sources/AF-ESC-2024, rating: very high)
- In the early post-PCI or ACS period, combined DOAC + P2Y12 inhibitor (dual antithrombotic therapy) is the recommended strategy for 6–12 months to balance stent thrombosis and stroke risk vs bleeding. (sources/ACS-AHA-2025, rating: very high)
- After the early period, the risk of stent thrombosis and recurrent ischemic events decreases. At this point, whether continuing the antiplatelet provides meaningful ischemic benefit over the bleeding cost is the key question. (sources/edoxaban-af-cad-nejm-2024, rating: high)
Foundational Trial Evidence Establishing Dual Over Triple Therapy (2018 EHRA)
Early RCTs establishing that dual therapy (OAC + P2Y12 without aspirin) reduces bleeding vs triple therapy without significant ischemic excess — the framework the 2018 EHRA guide codified:
- PIONEER AF-PCI (rivaroxaban): Rivaroxaban 15mg OD + P2Y12 vs VKA triple therapy in AF + PCI — both rivaroxaban arms significantly reduced clinically significant bleeding; non-standard rivaroxaban dose; underpowered for ischemic efficacy (sources/noac-ehra-2018, rating: high)
- RE-DUAL PCI (dabigatran 110mg or 150mg BID + P2Y12): Dual therapy vs VKA triple therapy — both dabigatran arms reduced major or CRNM bleeding; non-inferior for composite ischemic endpoint; dabigatran 150mg non-inferior for stroke prevention vs warfarin (sources/noac-ehra-2018, rating: high)
- 2018 EHRA recommendation: Triple therapy as short as possible (≤1 week); then OAC + P2Y12; no switch to VKA post-PCI (sources/noac-ehra-2018, rating: high)
Prior Evidence — OAC-ALONE and AFIRE Trials
- OAC-ALONE (2019, warfarin + antiplatelet vs OAC alone): Japanese RCT comparing warfarin monotherapy vs warfarin + single antiplatelet in AF patients >1 year after coronary stent implantation. Terminated prematurely with only 696 of planned 2000 patients; noninferiority of OAC alone was not demonstrated (HR for MACE 1.16; 95% CI 0.79–1.72); inconclusive. (sources/CCS-AHA-2023, rating: very high)
- AFIRE (2019, rivaroxaban): Rivaroxaban monotherapy vs rivaroxaban + antiplatelet in AF + stable CAD (>1 year post-PCI or confirmed anatomic CAD) — rivaroxaban monotherapy was noninferior for ischemic events (HR 0.59; 95% CI 0.39–0.89) and superior for MACE and bleeding; however predominantly low-risk patients enrolled; used non-standard rivaroxaban dose (15 mg or 10 mg once daily; Japan-specific approved dose, not global standard); terminated early for efficacy. Limits generalizability. (sources/CCS-AHA-2023, rating: very high; sources/edoxaban-af-cad-nejm-2024, rating: high)
EPIC-CAD Trial — Edoxaban Monotherapy (2024, NEJM)
The most definitive RCT to date, using standard-dose edoxaban in a broader stable CAD population.
Design: Multicenter, open-label, adjudicator-masked RCT; 1,040 patients; 18 South Korean sites; 12-month follow-up.
Population: AF (CHA₂DS₂-VASc ≥2) + stable CAD (PCI ≥6 months, or ACS-PCI ≥12 months, or medically managed ≥50% stenosis); mean age 72; mean CHA₂DS₂-VASc 4.3; mean HAS-BLED 2.2; 65.7% prior revascularization + 34.3% medically managed CAD.
Intervention: Edoxaban 60/30 mg daily (dose-reduced criteria) vs edoxaban + aspirin (61.8%) or clopidogrel (37.8%).
Results:
- Primary NACE at 12 months: 6.8% vs 16.2%; HR 0.44 (95% CI 0.30–0.65; P<0.001); NNT = 10.6
- Major ischemic events: 1.6% vs 1.8%; HR 1.23 (95% CI 0.48–3.10); NS — statistically similar
- Major + CRNM bleeding: 4.7% vs 14.2%; HR 0.34 (95% CI 0.22–0.53)
- Major bleeding: 1.3% vs 4.5%; HR 0.32 (95% CI 0.14–0.73)
- Benefit consistent across all subgroups including medically managed vs revascularized CAD, aspirin vs P2Y12 as antiplatelet, and edoxaban dose
Key message: OAC monotherapy markedly reduces net adverse events, driven overwhelmingly by bleeding reduction. Ischemic events appear similar, but the trial was not powered to detect ischemic differences — a small excess ischemic risk from dropping the antiplatelet cannot be excluded. (sources/edoxaban-af-cad-nejm-2024, rating: high)
ADAPT AF-DES — The DES-Specific Trial (2026, NEJM)
The first monotherapy trial restricted to patients with a second- or third-generation drug-eluting stent at least one year earlier, and the first to use clopidogrel rather than aspirin as the comparator antiplatelet.
Design: investigator-initiated, open-label, non-inferiority RCT with a prespecified sequential superiority test; 32 centres in South Korea; 960 patients; 12-month follow-up; non-inferiority margin 3.0 percentage points. (sources/adapt-af-des-nejm-2026, rating: high)
Population: AF with CHA₂DS₂-VASc ≥2; PCI ≥1 year earlier with a second- or third-generation DES; mean age 71.1 years; 21.4% women; mean CHA₂DS₂-VASc 4.1; mean HAS-BLED 2.7; median 32.8 months from PCI to randomisation; 64.8% had ACS at the index PCI.
Intervention: apixaban 5 mg twice daily or rivaroxaban 20 mg daily alone, versus apixaban 5 mg twice daily or rivaroxaban 15 mg daily plus clopidogrel 75 mg daily. Only apixaban and rivaroxaban were used.
| Outcome at 12 months | NOAC alone | NOAC + clopidogrel | Effect |
|---|---|---|---|
| Net adverse clinical events | 9.6% | 17.2% | RD −7.6 pp (−11.9 to −3.3); HR 0.54 (0.37–0.77); non-inferior and superior |
| Major or CRNM bleeding | 5.2% | 13.2% | HR 0.38 (0.24–0.60) |
| Major bleeding | 2.3% | 6.1% | HR 0.37 (0.18–0.74) |
| Stent thrombosis | 0.4% | 0.4% | HR 0.99 (0.14–7.02) |
| Stroke | 1.0% | 0.8% | HR 1.24 (0.33–4.64) |
| Ischaemic composite (CV death, MI, stent thrombosis, stroke, embolism) | 3.3% | 4.0% | HR 0.84 (0.43–1.63) |
(Table: sources/adapt-af-des-nejm-2026)
Key message: same structure as EPIC-CAD — the net benefit is a bleeding benefit, with no stent-thrombosis penalty from dropping clopidogrel beyond one year. A striking secondary observation: inappropriate NOAC underdosing occurred in 11.3% of the monotherapy group but 29.0% of the combination group, evidence that clinicians were offsetting the bleeding risk of continued antiplatelet therapy by weakening the anticoagulant. (sources/adapt-af-des-nejm-2026)
Mechanism of Net Benefit
- The bleeding rate with dual antithrombotic therapy (~14%) substantially exceeds the ischemic event rate (~2%) in stable CAD + AF patients. A NACE endpoint will therefore be dominated by bleeding, mathematically favouring the less potent antithrombotic. This is a structural limitation of NACE endpoints. (sources/edoxaban-af-cad-nejm-2024, rating: high)
- OAC alone is biologically plausible as sufficient therapy: OAC prevents cardioembolic stroke and reduces the thrombin-dependent component of coronary thrombosis; antiplatelet therapy targets ADP/thromboxane pathways, which are most relevant in the immediate post-stent period. Beyond 12 months, these additional pathways contribute less incrementally. (sources/CCS-AHA-2023, rating: very high)
Guideline Position After EPIC-CAD and ADAPT AF-DES
- 2023 AHA/ACC CCD Guideline: After 1 year post-PCI in patients with AF, DOAC monotherapy may be considered (COR 2b/C-LD) — a weak recommendation based on AFIRE. (sources/CCS-AHA-2023, rating: very high)
- EPIC-CAD and ADAPT AF-DES together give randomised support in the two populations the guidelines describe — stable CAD whether revascularised or medically managed, and DES-treated patients beyond one year — with consistent bleeding-driven net benefit. Guideline revision to a higher class is now supported by two independent trials, though neither has been incorporated into a published guideline. (sources/edoxaban-af-cad-nejm-2024; sources/adapt-af-des-nejm-2026)
- Neither trial supports stopping anticoagulation in favour of antiplatelet therapy, and neither tested dropping the antiplatelet earlier than 6–12 months.
Practical Guidance (Synthesis of Current Evidence)
- In stable AF + CAD patients beyond 6–12 months from the index event/PCI: OAC monotherapy with a DOAC is a reasonable strategy, particularly when CHA₂DS₂-VASc is elevated and HAS-BLED confers high bleeding risk.
- Both aspirin and P2Y12 inhibitors appear to have similar bleeding excess when combined with OAC in this stable phase.
- The antiplatelet component appears to add little ischemic protection in the chronic stable CAD phase when OAC is already on board.
- In DES-treated patients beyond one year, the antiplatelet can be stopped without a stent-thrombosis penalty. Stent thrombosis was 0.4% in both arms of ADAPT AF-DES, and the trial used clopidogrel as the comparator — the contemporary post-PCI antiplatelet. (sources/adapt-af-des-nejm-2026)
- Check the anticoagulant dose before accepting a bleeding signal. Underdosing was nearly three times more common when clopidogrel was continued; some of the excess bleeding in combination arms may be a prescribing behaviour rather than an intrinsic drug effect. (sources/adapt-af-des-nejm-2026)
- East Asian populations may have higher innate bleeding risk than Western populations — the absolute magnitude of benefit may be smaller in Western populations, but the directional effect is expected to be the same.
Contradictions / Open Questions
- EPIC-CAD underpowered for ischemic events: The trial was not designed to detect differences in MI, stroke, or CV death individually. HR 1.23 for ischemic events (95% CI 0.48–3.10) leaves open the possibility of a small but clinically meaningful excess ischemic risk with OAC monotherapy. All available trials lack statistical power to exclude this definitively. (sources/edoxaban-af-cad-nejm-2024, rating: high)
- East Asian–only population: East Asians have a higher propensity for bleeding on antithrombotic therapy and a somewhat lower absolute ischemic event rate vs Western populations. The absolute magnitude of the bleeding reduction benefit from monotherapy may be smaller in Western cohorts. (sources/edoxaban-af-cad-nejm-2024, rating: high)
- Net clinical outcome endpoint bias: NACE inherently favours the less potent antithrombotic when bleeding events are far more frequent than ischemic events. EPIC-CAD, AFIRE, and OAC-ALONE all used composite endpoints that include bleeding, which by design will reach statistical significance more easily than a pure ischemic endpoint. (sources/edoxaban-af-cad-nejm-2024, rating: high)
- AFIRE non-standard dose: AFIRE used rivaroxaban 10–15 mg vs the globally approved 20 mg/15 mg (renal) dose — the lower Japanese dose reduces bleeding but may also reduce ischemic protection. Direct comparability with EPIC-CAD (standard edoxaban) is limited. (sources/CCS-AHA-2023, rating: very high)
- Optimal antithrombotic timing — when to transition from dual to mono: RCTs have tested the ≥6–12 months threshold (post-PCI or post-ACS). Whether transitioning earlier (e.g., at 3 months) is safe is untested. (sources/CCS-AHA-2023, rating: very high)
- A second positive trial does not add ischaemic precision. ADAPT AF-DES points the same way as EPIC-CAD, but stroke was numerically higher with monotherapy (5 versus 4 events) and the ischaemic composite CI spanned unity; the possibility of a small ischaemic cost is unresolved. (sources/adapt-af-des-nejm-2026)
- Both trials are open-label, single-country (South Korean) and East Asian, where bleeding risk on antithrombotic therapy is higher; the absolute benefit in Western cohorts may be smaller. (sources/adapt-af-des-nejm-2026; sources/edoxaban-af-cad-nejm-2024)
- ADAPT AF-DES does not answer the high-bleeding-risk patient. Prior intracranial haemorrhage, recent gastrointestinal bleeding, severe anaemia or thrombocytopenia and high bleeding risk were exclusions — the group in which dropping clopidogrel is most attractive. (sources/adapt-af-des-nejm-2026)
Connections
- Related to entities/Atrial-Fibrillation — OAC indication; ACS and stable CAD antithrombotic sections
- Related to entities/Chronic-Coronary-Disease — stable CAD antithrombotic management
- Related to concepts/DAPT-Strategies — early post-PCI/ACS period vs stable phase; AFIRE/EPIC-CAD context
- Related to concepts/CHA2DS2-VA — stroke risk scoring for AF
- Related to entities/Acute-Coronary-Syndrome — early post-ACS antithrombotic period (different clinical phase)