Half-Dose Rivaroxaban versus Antiplatelet Therapy for Preventing Silent Cerebral Embolic Lesions after Left Atrial Appendage Occlusion: The HALO-SCE Trial
Authors, Journal, Affiliations, Type, DOI
- Kexin Wang, Linsheng Shi and Zhongbao Ruan (co-first authors), Minglong Chen and Weizhu Ju (co-corresponding authors); First Affiliated Hospital with Nanjing Medical University, Affiliated Hospital of Nantong University and Affiliated Taizhou People's Hospital.
- Three centres in mainland China.
- European Heart Journal, 2026, published online; DOI 10.1093/eurheartj/ehag734.
- Type: investigator-initiated, prospective, multicentre, open-label randomized controlled trial with blinded outcome assessment.
- ClinicalTrials.gov: NCT05671276.
- Supported by the Jiangsu Commission of Health and the Noncommunicable Chronic Diseases National Science and Technology Major Project.
Overview
HALO-SCE randomized 164 patients with atrial fibrillation who had successful LAAO confirmed at 45 days — no device-related thrombus and no peri-device leak ≥3 mm — to half-dose rivaroxaban 10 mg daily or antiplatelet therapy. Serial diffusion-weighted MRI and cognitive testing over 12 months showed new silent cerebral embolic lesions in 12.2% with rivaroxaban versus 31.7% with antiplatelet therapy (P=0.005). Cognitive scores diverged progressively in favour of rivaroxaban, with 12-month MMSE and MoCA differences of 2.56 and 2.67 points. The composite of death, clinical thromboembolism and major bleeding was numerically lower with rivaroxaban (2.4% versus 11.0%; P=0.057), while major bleeding occurred in none of the rivaroxaban patients and 3.7% of the antiplatelet group. The trial is small and exploratory for clinical endpoints but directly challenges the common practice of switching to antiplatelet therapy after LAAO.
Keywords
Left atrial appendage occlusion, atrial fibrillation, half-dose rivaroxaban, antiplatelet therapy, silent cerebral embolic lesions, diffusion-weighted MRI, cognitive function, Watchman, randomized controlled trial
Key Takeaways
Background and Rationale
- The optimal long-term antithrombotic strategy after successful LAAO is undefined.
- Even after LAAO, residual thromboembolic risk persists from blood stasis, endothelial dysfunction and hypercoagulability in the fibrillating atrium; antiplatelet therapy does not address these mechanisms.
- Silent cerebral embolic lesions occur after LAAO and have been associated with subsequent cognitive decline.
- The trial tested whether half-dose rivaroxaban reduces silent cerebral embolic lesions and preserves cognition compared with antiplatelet therapy.
Methods
- Patients with AF who underwent percutaneous LAAO with Watchman or Watchman FLX between December 2022 and February 2025 were screened at three Chinese centres.
- Enrolment after the 45-day evaluation confirmed no device-related thrombus and no significant peri-device leak (residual leak <3 mm) on cardiac CT or transesophageal echocardiography.
- All patients received rivaroxaban 15 mg plus aspirin 100 mg immediately after LAAO until the 45-day evaluation.
- Randomized 1:1 to rivaroxaban 10 mg daily until at least 12 months, or aspirin 100 mg plus clopidogrel 75 mg daily until 180 days followed by aspirin 100 mg daily through 365 days.
- Diffusion-weighted MRI and cognitive assessment (MMSE and Chinese MoCA) at baseline, 90, 180 and 365 days.
- Primary outcome: patient-level incidence of any newly detected silent cerebral embolic lesion on follow-up diffusion-weighted MRI within 1 year.
- Secondary outcomes: cognitive performance, proportion with ≥2 new lesions, cumulative lesion number and size, and a composite of all-cause mortality, clinical thromboembolic events and major bleeding.
Participants
- 177 screened, 164 randomized: 82 to half-dose rivaroxaban and 82 to antiplatelet therapy.
- Evaluable follow-up MRI in 80 and 78 patients; evaluable cognitive assessments in 80 and 80.
- Mean age 69.2±7.1 years; 36.6% women; 77.4% had prior stroke.
- Mean CHA₂DS₂-VASc 4.2±1.3; mean HAS-BLED 2.4±0.7.
- Baseline MMSE 24.5±3.9; baseline MoCA 19.6±4.4.
- 72.0% and 74.4% underwent LAAO combined with ablation; Watchman FLX in 6.1% of each group.
Primary Imaging Outcome
- New silent cerebral embolic lesions: 10/82 (12.2%) with rivaroxaban versus 26/82 (31.7%) with antiplatelet therapy (P=0.005).
- ≥2 new lesions: 9/82 (11.0%) versus 21/82 (25.6%; P=0.015).
- Cumulative lesion distribution differed significantly (P=0.003); ≥5 lesions occurred in 0% versus 11.0%.
- Cumulative lesion size >10 mm: 0% versus 20.7% (P<0.001).
- Findings were consistent in the evaluable-MRI and per-protocol analyses.
Cognitive Outcomes
- Cognitive decline (≥2-point MoCA decrease) occurred in 32/160 (20.0%) overall: 4/80 (5.0%) with rivaroxaban versus 28/80 (35.0%) with antiplatelet therapy (P<0.001).
- Baseline MMSE and MoCA did not differ significantly; trajectories diverged progressively.
- At 365 days, model-derived between-group differences favoured rivaroxaban: MMSE 2.56 points (95% CI 1.11–4.01; P<0.001) and MoCA 2.67 points (95% CI 1.07–4.26; P=0.001).
- Benefits were consistent regardless of concomitant ablation.
Clinical Outcomes and Adherence
- Composite of all-cause mortality, clinical thromboembolic events and major bleeding: 2.4% versus 11.0% (P=0.057).
- Clinical thromboembolic events: 1.2% versus 7.3% (P=0.117).
- Major bleeding: 0% versus 3.7% (P=0.246).
- Treatment adherence through 12 months: 95.1% versus 85.4% (P=0.035).
Interpretation
- Half-dose rivaroxaban after successful LAAO reduced silent cerebral embolic lesions and was associated with more favourable cognitive trajectories than antiplatelet therapy.
- The findings support a mechanistically targeted maintenance strategy in patients who are eligible for oral anticoagulation and have confirmed successful occlusion without device thrombus or significant leak.
- The results differ from BRAIN-AF, which tested low-dose rivaroxaban in a lower-risk AF population and did not reduce a composite of cognitive decline, stroke and TIA; HALO-SCE enrolled a higher-risk population with higher CHA₂DS₂-VASc scores and more prior stroke.
- Whether half-dose anticoagulation should be continued lifelong after LAAO, or withdrawn, remains uncertain.
Limitations of the Document
- Small sample size (n=164) powered primarily for the imaging endpoint; clinical outcomes are exploratory and underpowered.
- Open-label treatment allocation; MRI and cognitive assessments were blinded, but clinical management was not.
- The absence of major bleeding in the rivaroxaban arm cannot establish absolute safety.
- Detailed sociodemographic and neuropsychological baseline parameters were not routinely collected.
- Results apply only to patients with successful Watchman or Watchman FLX LAAO, no significant leak or device thrombus, and no early major bleeding or thromboembolic event.
- The findings are limited to rivaroxaban 10 mg and cannot be extrapolated to other anticoagulants or doses.
- All patients received the same initial regimen before randomization, so the trial provides no evidence on the optimal therapy during the first 45 days after LAAO.
- Single-country study, which limits generalizability.
Key Concepts Mentioned
- concepts/NOAC-AF-Management — anticoagulant dosing and post-procedural management
- concepts/CHA2DS2-VA — stroke risk stratification in AF
- concepts/Subclinical-AF — the broader concept of silent atrial arrhythmia and embolic risk
- concepts/Poststroke-AF-Monitoring — detection of atrial fibrillation after stroke
Key Entities Mentioned
- entities/LAAO — the device procedure and post-implant antithrombotic strategy
- entities/Atrial-Fibrillation — the underlying arrhythmia
- entities/Ischemic-Stroke — the clinical thromboembolic outcome
Wiki Pages Updated
wiki/sources/Half-NOAC-LAAO-EHJ-2026.md— created (this file)wiki/entities/LAAO.md— updated with HALO-SCE and the post-implant antithrombotic debatewiki/concepts/NOAC-AF-Management.md— updated with post-LAAO half-dose evidencewiki/entities/Atrial-Fibrillation.md— updated with the cognitive/embolic linkwiki/wikiindex.md— updatedwiki/sourceindex.md— updated