← Back
NeuroTrials.ai
Neurology Clinical Trial Database

STABLED

Stroke Secondary Prevention With Catheter Ablation and Edoxaban for Patients With Nonvalvular Atrial Fibrillation

Year of Publication: 2026

Authors: Kazumi Kimura, Yasuhiro Nishiyama, Yu-ki Iwasaki, ..., Ken Okumura

Journal: JAMA Neurology

Citation: Kimura K et al. Catheter Ablation and Oral Anticoagulation for Secondary Stroke Prevention in Atrial Fibrillation: The STABLED Randomized Clinical Trial. JAMA Neurol. Published online March 2, 2026. doi:10.1001/jamaneurol.2026.0155

Link: https://jamanetwork.com/journals/jamaneu...eurol.2026.0155


Clinical Question

Does adding catheter ablation to standard edoxaban therapy reduce the risk of recurrent ischemic stroke or composite cardiovascular outcomes in patients with nonvalvular atrial fibrillation and a recent history of ischemic stroke?

Bottom Line

In patients with nonvalvular AF and a recent ischemic stroke, catheter ablation added to edoxaban did not significantly reduce the primary composite endpoint of recurrent ischemic stroke, systemic embolism, all-cause death, or hospitalization for heart failure compared to edoxaban alone. The trial was likely underpowered due to lower-than-anticipated event rates.

Major Points

  • Multicenter open-label RCT conducted at 45 sites in Japan; 249 patients randomized 1:1
  • Median follow-up exceeded 3 years (3.7 years standard therapy, 3.5 years ablation)
  • Primary composite endpoint occurred at identical rates per group (18.0% each) but with slightly higher per-person-year rate in ablation arm (5.6 vs 4.9 per 100 person-years; HR 1.11, P=.70)
  • Recurrent ischemic stroke trended numerically lower with ablation (2.5 vs 3.1 per 100 person-years; HR 0.75) but did not reach significance
  • All-cause mortality was numerically higher in the ablation arm (2.8 vs 1.0 per 100 person-years; HR 2.79; P=.05), though confidence intervals crossed 1.0
  • Hospitalization for heart failure was numerically lower with ablation (1.0 vs 1.5 per 100 person-years; HR 0.66) but not significant
  • Major bleeding was more frequent in the ablation arm (1.9 vs 0.6 per 100 person-years; HR 2.99), also not significant
  • Two ablation-related serious adverse events: 1 cardiac tamponade and 1 stroke (0.8% each)
  • Crossover rate was 14% overall (15.2% ablation-to-control; 12.9% control-to-ablation)
  • Study was underpowered: assumed 15% annual event rate in control arm, actual rate was approximately 4.9 per 100 person-years
  • Conducted exclusively in Japan; generalizability to other ethnic populations uncertain

Design

Study Type: Multicenter open-label parallel-group randomized clinical trial

Randomization: 1

Blinding: Open-label; no blinding of investigators or participants. Randomization stratified by CHADS2 score (<4 vs >=4) and AF type (paroxysmal vs other) via web-based allocation system (EPS Corporation).

Enrollment Period: January 2018 to March 2021

Follow-up Duration: Minimum 3 years from final patient randomization; maximum approximately 6 years; median 3.7 years (standard therapy) and 3.5 years (ablation)

Centers: 45

Countries: Japan

Sample Size: 249

Analysis: Intention-to-treat (ITT) primary analysis; per-protocol set as sensitivity analysis. Log-rank test for primary endpoint P-value; Kaplan-Meier curves for time-to-first-event. Hazard ratios estimated via multivariate Cox proportional hazards model with CHADS2 score and AF type as covariates. Exploratory analysis using time-dependent covariate Cox model for ablation performed vs not performed. No imputation of missing data. SAS version 9.4.


Inclusion Criteria

  • Age 20-85 years
  • Definitive diagnosis of nonvalvular atrial fibrillation on ECG
  • History of ischemic stroke within 6 months prior to enrollment
  • Currently receiving or scheduled to receive edoxaban
  • Modified Rankin Scale score of 3 or less
  • Written informed consent provided

Exclusion Criteria

  • Age <20 or >85 years
  • Valvular atrial fibrillation
  • Modified Rankin Scale score >3
  • Severe stroke (implied by mRS exclusion and DWI-ASPECTS considerations)
  • Active cancer (implied exclusion based on discussion section reference)
  • Left ventricular ejection fraction <=35% (implied exclusion based on discussion reference to CASTLE-AF population differences)
  • Creatinine clearance <30 mL/min (no patients enrolled with CrCl <30)
  • Protocol-defined Good Clinical Practice/ICH violations (2 patients excluded post-enrollment for consent withdrawal)

Arms

FieldControlStandard Therapy + Catheter Ablation
InterventionEdoxaban 60 mg once daily (or 30 mg for patients meeting dose-adjustment criteria: weight <=60 kg, CrCl 15-50 mL/min, or concomitant P-gp inhibitor use). Antiarrhythmic drugs permitted as needed under cardiologist supervision.Edoxaban (same dosing as control) plus catheter ablation. Ablation performed after at least 4 weeks of edoxaban therapy and within 1-6 months of index stroke onset. Technique: pulmonary vein isolation with additional atrial ablation at operator discretion. Repeat ablation permitted (18 patients, 14.4%, underwent repeat procedure).
DurationMinimum 3 years; maximum approximately 6 yearsAblation performed at median 65 days post-randomization (range 10-442 days); followed minimum 3 years

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Composite of recurrent ischemic stroke, systemic embolism, all-cause death, and hospitalization for heart failure (time to first event)Primary22 events (18.0%); 4.9 per 100 person-years22 events (18.0%); 5.6 per 100 person-years1.11.70
Recurrent ischemic strokeSecondary14 events (11.5%); 3.1 per 100 person-years10 events (8.2%); 2.5 per 100 person-years0.75 (95% CI 0.33-1.70).49
Hemorrhagic strokeSecondary0 events3 events (2.5%); 0.7 per 100 person-yearsNA
Systemic embolismSecondary0 events0 eventsNANA
All-cause deathSecondary5 events (4.1%); 1.0 per 100 person-years12 events (9.8%); 2.8 per 100 person-years2.79 (95% CI 0.98-7.92).05
Cardiovascular deathSecondary2 events (1.6%); 0.4 per 100 person-years6 events (4.9%); 1.4 per 100 person-years3.47 (95% CI 0.70-17.2)
Hospitalization for heart failureSecondary7 events (5.7%); 1.5 per 100 person-years4 events (3.3%); 1.0 per 100 person-years0.66 (95% CI 0.19-2.27).51
Major bleeding (ISTH criteria)Secondary3 events (2.5%); 0.6 per 100 person-years8 events (6.6%); 1.9 per 100 person-years2.99 (95% CI 0.79-11.3)
Intracranial hemorrhageSecondary2 events (1.6%); 0.4 per 100 person-years4 events (3.3%); 0.9 per 100 person-years2.26 (95% CI 0.41-12.4)
Gastrointestinal bleedingSecondary13 events (10.7%); 2.9 per 100 person-years10 events (8.2%); 2.4 per 100 person-years0.82 (95% CI 0.36-1.88)
Net clinical outcome (composite of all-cause death, systemic embolism, stroke, hospitalization for heart failure, and serious ablation-related adverse events)Secondary22 events (18.0%); 4.9 per 100 person-years23 events (18.9%); 5.9 per 100 person-years1.18 (95% CI 0.66-2.11)
Cardiac tamponade (ablation-related)Adverse01 (0.8%)
Stroke (ablation-related)Adverse01 (0.8%)

Subgroup Analysis

Exploratory Cox model with catheter ablation as a time-dependent covariate was performed to account for crossover and pre-procedure events. This analysis yielded a numerically lower primary endpoint incidence rate in the ablation-performed group (HR 0.84; 95% CI 0.46-1.56), though still not statistically significant. No association was found between timing of catheter ablation and time to events. Per-protocol set analyses for all key endpoints were consistent with ITT results.


Criticisms

  • Open-label design introduces potential performance and detection bias
  • Severely underpowered: actual control event rate (~4.9/100 person-years) was far lower than the assumed 15% annual rate used for power calculations, leaving the trial unable to detect clinically meaningful differences
  • Crossover rate of 14% dilutes treatment effect estimates in ITT analysis, consistent with the CABANA trial paradox
  • Pre-procedure events in the ablation arm (2 patients experienced primary endpoint before ablation) may have inflated apparent harm in the ablation group
  • Trial conducted exclusively in Japan (all-Japanese population), limiting generalizability to other ethnic groups
  • COVID-19 pandemic-related visitation/operational restrictions may have limited cause-of-death ascertainment
  • Ablation group had higher baseline antiplatelet use (14.4% vs 4.0%) and more frequent heart disease history (19.2% vs 14.5%), potentially confounding mortality and bleeding outcomes
  • Higher CHA2DS2-VASc scores in ablation arm may reflect selection or stratification imbalance
  • Sample size (249) was small for a complex cardiovascular endpoint trial with median 3+ year follow-up
  • Edoxaban was the sole anticoagulant; results may not generalize to other DOACs
  • No sham procedure control; open-label ablation may influence concomitant medication use and follow-up intensity

Funding

Daiichi Sankyo Co, Ltd (manufacturer of edoxaban). Three authors (Fukuzawa, Sugimoto, Takita) were Daiichi Sankyo employees. Funder involved in study design, data analysis planning, interpretation, and publication decision.

Based on: STABLED (JAMA Neurology, 2026)

Authors: Kazumi Kimura, Yasuhiro Nishiyama, Yu-ki Iwasaki, ..., Ken Okumura

Citation: Kimura K et al. Catheter Ablation and Oral Anticoagulation for Secondary Stroke Prevention in Atrial Fibrillation: The STABLED Randomized Clinical Trial. JAMA Neurol. Published online March 2, 2026. doi:10.1001/jamaneurol.2026.0155

Content summarized and formatted by NeuroTrials.ai.