#atrial-fibrillation #drug-eluting-stent #anticoagulation #antiplatelet-therapy #NOAC #randomized-controlled-trial

Therapy for Atrial Fibrillation in Patients with Drug-Eluting Stents (ADAPT AF-DES)

Authors, Journal, Affiliations, Type, DOI

Authors: Seung-Jun Lee, Hee Tae Yu, Yong-Joon Lee, Sang-Hyup Lee, Jung Ho Heo, Sung Gyun Ahn, Sanghoon Shin, Joon-Hyung Doh, Ae-Young Her, Byung-Ryul Cho, Gwang-Sil Kim, Taek-Geun Kwon, Sang-Wook Lim, Jaemin Shim, Ji-Yong Jang, Kyounghoon Lee, Yun-Hyeong Cho, Cheol Ung Choi, Sang-Rok Lee, Hyung-Bok Park, Han Cheol Lee, Seunghwan Kim, Kyeong Ho Yun, Jong-Hwa Ahn, Byoung-Kwon Lee, Deok-Kyu Cho, Song-Yi Kim, Ung Kim, Tae Soo Kang, Seong Huan Choi, Won-Ho Kim, Jin Bae Lee, Moo-Yong Rhee, Jin-Bae Kim, Sang-Ho Jo, Dae Woo Hyun, Daehoon Kim, Tae-Hoon Kim, Sung-Jin Hong, Jae-Sun Uhm, Dong-Ho Shin, Chul-Min Ahn, Byeong-Keuk Kim, Boyoung Joung, Young-Guk Ko, Donghoon Choi, Myeong-Ki Hong, Yangsoo Jang, Hui-Nam Pak, Jung-Sun Kim, for the ADAPT AF-DES Investigators
Journal: New England Journal of Medicine, 2026;394(7):658-668 (published online November 8, 2025; updated January 6, 2026)
Affiliations: Severance Hospital, Yonsei University College of Medicine, Seoul, and 31 other centres in South Korea
Type: Investigator-initiated, multicentre, randomised, open-label, non-inferiority trial. Funded by the Cardiovascular Research Center (South Korea) under a contract with Samjin Pharmaceutical; the funders had no role in design, analysis, interpretation or writing.
DOI: https://doi.org/10.1056/NEJMoa2512091

Overview

ADAPT AF-DES tested whether a non-vitamin K antagonist oral anticoagulant (NOAC) alone is non-inferior to a NOAC plus clopidogrel in patients with atrial fibrillation who had received a second- or third-generation drug-eluting stent at least one year earlier. In 960 patients randomised at 32 South Korean centres, the 12-month net adverse clinical event rate was 9.6% with monotherapy and 17.2% with combination therapy — non-inferior on the prespecified 3.0-percentage-point margin and superior in the prespecified sequential test (HR 0.54). The effect was driven almost entirely by bleeding (major or clinically relevant non-major bleeding 5.2% versus 13.2%), while ischaemic events were similar. It is the first trial dedicated to stable patients beyond one year after drug-eluting stent implantation, and it aligns with AFIRE and EPIC-CAD.

Keywords

Atrial fibrillation; drug-eluting stent; percutaneous coronary intervention; anticoagulation; NOAC; clopidogrel; antithrombotic therapy; net adverse clinical events; bleeding

Key Takeaways

Background

AF requires long-term anticoagulation; PCI with drug-eluting stents (DES) requires antiplatelet therapy to prevent stent thrombosis. Combining them increases bleeding.
Guidelines recommend dropping aspirin early and continuing a NOAC plus a P2Y12 inhibitor for 6–12 months after PCI, then oral anticoagulant monotherapy beyond one year — but the randomised evidence for that final step was limited.
The three prior monotherapy trials — OAC-ALONE, AFIRE and EPIC-CAD — did not focus on DES, allowed warfarin or aspirin, or were terminated early. ADAPT AF-DES was designed to close that gap in a DES-specific population.

Methods

Design: multicentre, randomised, open-label, non-inferiority trial with a prespecified sequential superiority test, at 32 centres in South Korea, April 2020 to May 2024.
Population: age 19–85 years; AF with CHA₂DS₂-VASc ≥2; PCI with a second- or third-generation DES at least 1 year before enrolment. Exclusions included mechanical prosthetic valve, moderate-to-severe mitral stenosis, prior intracranial haemorrhage, recent gastrointestinal bleeding, high bleeding risk, severe anaemia or thrombocytopenia, active cancer, and first-generation DES.
Intervention: 1:1 to NOAC monotherapy (apixaban 5 mg twice daily or rivaroxaban 20 mg once daily) or combination therapy (apixaban 5 mg twice daily or rivaroxaban 15 mg once daily plus clopidogrel 75 mg daily), with prespecified dose reductions. Only apixaban and rivaroxaban were used in the trial.
Primary end point: net adverse clinical events at 12 months — death from any cause, myocardial infarction, stent thrombosis, stroke, systemic embolism, or major bleeding or clinically relevant non-major bleeding (ISTH definitions).
Statistics: non-inferiority margin 3.0 percentage points (upper bound of the one-sided 97.6% CI), 80% power, planned 960 patients; an O'Brien–Fleming interim analysis after one third of enrolment set the final one-sided alpha at 2.4%, so a 95.2% CI was used for the primary end point.

Results

Participants

1,283 assessed; 960 randomised (482 monotherapy, 478 combination); 99.4% and 99.2% completed 12-month follow-up.
Mean age 71.1 ± 8.5 years; 21.4% women; mean CHA₂DS₂-VASc 4.1; mean HAS-BLED 2.7–2.8; 64.8% had ACS at the index PCI; complex PCI features in 25.7%.
Median time from PCI to randomisation 32.8 months (IQR 15.1–73.5); 61.1% had paroxysmal AF.
Before randomisation, 61.1% were on combined antiplatelet plus anticoagulant therapy, 23.1% on anticoagulant alone, 15.7% on antiplatelet alone.
Inappropriate underdosing of apixaban or rivaroxaban occurred in 11.3% of the monotherapy group and 29.0% of the combination group — physicians' bleeding concern visible in prescribing.

Primary Outcome

46/482 (9.6%) monotherapy versus 82/478 (17.2%) combination therapy; absolute difference −7.6 percentage points (95.2% CI −11.9 to −3.3; P<0.001 for non-inferiority); HR 0.54 (95.2% CI 0.37–0.77; P<0.001 for superiority).
Per-protocol analysis consistent with the intention-to-treat result.

Secondary Outcomes at 12 Months

Outcome	NOAC alone	NOAC + clopidogrel	HR (95% CI)
Death from any cause	2.9%	4.0%	0.73 (0.37–1.46)
Cardiovascular death	1.5%	2.3%	0.63 (0.24–1.63)
Myocardial infarction	0.8%	1.3%	0.66 (0.19–2.34)
Stent thrombosis	0.4%	0.4%	0.99 (0.14–7.02)
Stroke	1.0%	0.8%	1.24 (0.33–4.64)
Systemic embolism	0	0	—
Major bleeding or CRNM bleeding	5.2%	13.2%	0.38 (0.24–0.60)
Major bleeding	2.3%	6.1%	0.37 (0.18–0.74)
Clinically relevant non-major bleeding	2.9%	7.1%	0.40 (0.21–0.74)
CV death, MI, stent thrombosis, stroke or systemic embolism	3.3%	4.0%	0.84 (0.43–1.63)

The net benefit is a bleeding benefit. The pure ischaemic composite was numerically but not significantly lower with monotherapy, and stroke was numerically higher (5 versus 4 events). The trial was not powered for individual ischaemic outcomes.

Subgroups

Consistent direction across prespecified subgroups, with larger apparent effects in patients ≥75 years (HR 0.41), women (HR 0.32), those ≥24 months from PCI (HR 0.45), those with CHA₂DS₂-VASc ≥4 (HR 0.46), and complex PCI (HR 0.42, post hoc).
Subgroup confidence intervals were not adjusted for multiplicity.

Discussion

The authors position ADAPT AF-DES as the DES-specific confirmation of AFIRE and EPIC-CAD, and note that the 2026 guideline position (anticoagulant monotherapy after one year) had been based on limited randomised evidence.
The high rate of inappropriate underdosing when clopidogrel was continued is presented as a real-world signal of clinicians trading anticoagulant intensity for bleeding safety — and as a mechanism that may have inflated the bleeding difference.
Prespecified secondary analyses with follow-up to 2 years are ongoing.

Limitations of the document

Open-label, although all clinical end points were adjudicated by an independent committee blinded to assignment.
Lower-than-anticipated event rates reduced statistical power; because the trial was designed for non-inferiority, the superiority finding should be interpreted cautiously.
Only apixaban and rivaroxaban were used, and the combination arm used a reduced rivaroxaban dose (15 mg) by design — extrapolation to edoxaban, dabigatran, or standard-dose rivaroxaban combination therapy is uncertain.
Inappropriate underdosing was three times more frequent in the combination arm, confounding the bleeding comparison.
Single country, East Asian population; the East Asian bleeding–ischaemia balance may differ from Western cohorts, so absolute benefit may be smaller elsewhere.
One-year follow-up may be too short for stent-related and bleeding outcomes in stable CAD.
Patients at high bleeding risk, with prior intracranial haemorrhage, or with first-generation stents were excluded — the group most likely to benefit from monotherapy may be under-represented.
Industry funding through a contract with Samjin Pharmaceutical, though with explicit firewall provisions.

Key Concepts Mentioned

Key Entities Mentioned

Wiki Pages Updated

log.md carries the ingest entry.