The key point: The CAAN-AF (CRT And AV Node ablation trial in AF), published in Heart Rhythm O2 (2025), is the first randomized clinical trial designed to answer whether atrioventricular (AV) node ablation is superior to pharmacological rate control in patients with cardiac resynchronization therapy with a defibrillator (CRT-D) and permanent atrial fibrillation (AF). The trial plans to enroll 590 participants, with a primary composite outcome of all-cause mortality and nonfatal heart failure (HF) events over 2 years. The hypothesis: by ensuring 100% biventricular pacing, AV node ablation may restore the benefit of CRT that is lost in the presence of AF.
Why this study is needed
Cardiac resynchronization therapy (CRT, short for cardiac resynchronization therapy) is an established treatment for patients with symptomatic heart failure (NYHA class II to IV), left ventricular ejection fraction (LVEF) at or below 35% and QRS duration at or above 120 ms. Trials such as COMPANION, CARE-HF and MADIT-CRT showed significant reductions in mortality and HF hospitalizations.
However, the main CRT clinical trials excluded patients with atrial fibrillation. Observational data are clear: patients with permanent AF do not obtain the same benefit from CRT as those in sinus rhythm. The RAFT study subanalysis (Resynchronization for Ambulatory Heart Failure Trial) confirmed this observation: there was no difference in the primary outcome of death or HF hospitalization between CRT-D and ICD alone in patients with permanent AF (HR 0.96; 95% CI 0.65-1.41; P=0.82).
The most widely accepted explanation concerns the percentage of biventricular pacing. In AF, irregular conduction through the AV node produces intrinsic beats that compete with CRT pacing. This produces three phenomena that compromise effectiveness: unpaced beats (the ventricle is activated by native conduction before CRT), fusion beats (native conduction and pacing occur simultaneously), and pseudofusion (the CRT impulse occurs after native conduction has already activated the ventricle). As a result, the percentage of actual biventricular pacing falls, even when the device counter records high values.
AV node ablation eliminates this problem. By destroying the AV conduction pathway, each heartbeat becomes entirely dependent on CRT biventricular pacing. Observational data and a meta-analysis published in the Journal of the American College of Cardiology showed that AV node ablation reduced all-cause mortality with a relative risk of 0.42 (95% CI 0.26-0.68; P<0.001) compared with pharmacological rate control. The trade-off: the patient becomes permanently dependent on a pacemaker.
Despite these data, the 2021 ESC guidelines on pacemakers and CRT give AV node ablation only a class IIa recommendation with level B evidence (for biventricular pacing below 90-95%) and class IIa with level C evidence for the general recommendation. The absence of a dedicated randomized trial prevents a class I recommendation. CAAN-AF was designed to fill this gap.
Key CAAN-AF design data
- 590 participants planned (295 per arm), 1:1 randomization
- Type: phase IV, international, multicenter, investigator-initiated, prospective
- Intervention: AV node ablation vs. pharmacological rate control (target HR <90 bpm at rest)
- Primary outcome: composite of all-cause mortality + nonfatal HF events over 2 years
- Statistical power: 80% to detect a 25% relative reduction in the primary outcome (two-sided alpha 5%)
- Estimated control event rate: 50 events per 100 person-years (based on COMPANION)
- Registration: NCT01522898
Study design
Inclusion criteria
CAAN-AF includes patients aged 18 years or older with permanent AF, symptomatic HF (NYHA class II, III or ambulatory class IV), documented LVEF at or below 35% by echocardiogram or cardiac MRI, and QRS duration at or above 120 ms on a 12-lead electrocardiogram. Permanent AF is defined as AF in which achieving and maintaining sinus rhythm is considered unfeasible or ineffective in the long term, or when the patient and physician accept the presence of AF and a rhythm control strategy is no longer pursued, or when sinus rhythm cannot be restored.
Exclusion criteria
Excluded are: people younger than 18 years, pregnant patients, patients with prior AV node ablation, second- or third-degree AV block, inability to provide consent, life expectancy under 24 months due to a comorbidity unrelated to HF (for example, cancer, end-stage kidney disease or liver failure), and paroxysmal AF that terminates spontaneously within 7 days.
Two enrollment pathways
The study provides two pathways. Pathway A is for patients without prior CRT-D: they undergo preimplant assessment, CRT-D implantation, optimization of HF therapy and device programming (in 30 to 180 days), baseline assessment and then randomization. Pathway B is for patients with CRT-D in place for more than 30 days: HF therapy and the device are optimized before baseline assessment and randomization.
Control arm
Pharmacological rate control at the treating physician's discretion. Permitted drugs include beta-blockers, calcium channel blockers, digoxin or amiodarone, alone or in combination. The resting heart rate target, in the absence of symptoms, is below 90 beats per minute.
Intervention arm
Radiofrequency AV node ablation via femoral vein puncture, with the catheter positioned in the compact AV node region. The procedural endpoint is complete AV block after a waiting period of at least 10 minutes after ablation. Ablation should occur as close as possible to randomization, within a maximum of 6 weeks.
Outcomes and statistical analysis
Primary outcome
Composite of all-cause mortality and nonfatal HF events. HF events are defined according to MADIT-CRT: signs and/or symptoms of congestive HF responsive to intravenous diuretics or inotropes in an outpatient or emergency setting (with or without admission), or intensification of oral/parenteral decongestion in patients already hospitalized.
Secondary outcomes
Secondary outcomes include: all-cause mortality, cardiovascular mortality, nonfatal HF events, unplanned hospitalizations, ventricular arrhythmias requiring device therapy, inappropriate ICD shocks, 6-minute walk test, quality of life (SF-36, Minnesota Living with Heart Failure and CES-Depression questionnaires), left ventricular end-systolic diameter, percentage of biventricular pacing, reverse ventricular remodeling and NT-proBNP.
Sample size calculation and interim analyses
The calculation began with observational data showing a 58% reduction in mortality with AV node ablation. Conservatively, the study was powered to detect a 25% relative reduction in the composite outcome. With an event rate of 50 per 100 person-years in the control group (based on COMPANION), 80% power, two-sided alpha of 5% and a 10% allowance for withdrawal or crossover, 295 participants per arm are needed (590 total). Crossovers are allowed only after a patient reaches the primary outcome.
The design includes three formal interim analyses at 25%, 50% and 75% of accumulated information, following O'Brien-Fleming boundaries. P values for stopping for efficacy are: below 0.000015 at the first, 0.0031 at the second and 0.014 at the third analysis, with P below 0.044 for the final analysis. Futility analyses accompany each checkpoint. The primary analysis will be conducted by intention to treat, using negative binomial regression adjusted for demographic parameters and risk factors.
Implications for clinical practice
CAAN-AF addresses a real gap in cardiology: approximately 25% of patients with CRT develop AF, and current guidelines base the AV node ablation recommendation on observational data. The 58% reduction in mortality reported in the meta-analysis is impressive, but it comes from nonrandomized studies with inherent selection bias. By randomizing 590 patients with adequate power and prespecified interim analyses, CAAN-AF has the potential to transform a IIa/B recommendation into I/A.
An important aspect of the design is the requirement to optimize HF therapy and device programming before randomization. This ensures that any benefit observed with AV node ablation is incremental to drug treatment and optimized CRT programming, rather than simply reflecting more attentive care.
My perspective as an electrophysiologist: AV node ablation in patients with CRT and AF is a decision I regularly face in practice. The physiological rationale is sound: ensuring actual biventricular pacing, not merely pacing counted by the device. However, making a patient permanently dependent on a pacemaker is an irreversible decision that requires robust evidence. CAAN-AF is positioned to provide this answer. While awaiting results, I use AV node ablation selectively, mainly when actual biventricular pacing remains below 90-95% despite optimization of medication and device programming, in line with the current ESC recommendation.
