Catheter Ablation for Atrial Fibrillation

Contents. #Definition · #Indications and patient selection · #Technique and energy sources · #Transseptal access after septal closure · #Procedural complications · #Lesion sets and targets · #Periprocedural anticoagulation · #Surgical and hybrid ablation · #Outcomes beyond arrhythmia control · #Genetic determinants of outcome · #Contradictions / Open Questions · #Sources

Definition

Catheter ablation for AF (primarily pulmonary vein isolation, PVI) is a percutaneous interventional rhythm control strategy that electrically isolates the pulmonary veins — the dominant source of AF triggers — from the left atrial body. It is now the most evidence-based rhythm control intervention for AF, particularly for paroxysmal AF, with an expanded role in persistent AF and selected patients with HFrEF.

Key Concepts

Indications and patient selection

• ESC first-line upgrade. Class I/A for paroxysmal AF within a shared decision-making rhythm control strategy, reflecting consistent RCT evidence of better symptom control, fewer recurrences and lower AF burden than AADs; first-line ablation may be chosen after shared discussion. (sources/AF-ESC-2024, rating: very high)
• HFrEF, ESC. Class I when tachycardia-induced cardiomyopathy is likely — curative in that subgroup — and IIa in selected other HFrEF patients to improve prognosis (CASTLE-AF: reduced all-cause mortality and HF hospitalisation). CABANA found no significant mortality or morbidity benefit in its primary analysis, with benefit confined to per-protocol subgroups. (sources/AF-ESC-2024)
• HFrEF, AHA. Class I/A when GDMT is optimised and benefit is reasonably expected: CASTLE-AF composite of death or HF hospitalisation HR 0.62; CABANA HF substudy 46% mortality reduction. Suspected tachycardiomyopathy warrants aggressive early rhythm control, with ablation strongly preferred. (sources/AF-AHA-2023, rating: very high)
• First-line ablation prevents progression. In younger patients with few comorbidities, first-line ablation reduced the development of persistent AF versus AADs (HR 0.25). (sources/AF-AHA-2023)
• EAST-AFNET 4 context. Ablation was used in only ~8% of patients initially, rising to 19.4% at 2 years, so most of the early rhythm-control benefit was attributable to AADs — the strategy effect is not exclusively an ablation effect. (sources/EAST-AFNET4-NEJM-2020, rating: very high)
• Persistent AF is the weak link. Technique varies widely and the optimal strategy is unknown. (sources/AF-ESC-2024)

AVANT GUARD — first-line PFA in persistent AF.

• International multicentre RCT (n=310, previously untreated persistent AF): PFA (PVI plus posterior wall, pentaspline FARAWAVE) versus AADs (flecainide 57%, sotalol 28%), all patients monitored with an implantable cardiac monitor.
• Treatment success at 12 months 56% versus 30% (HR 0.46; P<0.001); median AF burden 0% versus 0.2%.
• The first RCT establishing first-line ablation as superior to AADs in persistent AF, extending the first-line paradigm beyond paroxysmal disease.
• Six procedure-related strokes (2%) paused the trial and forced protocol changes — mandatory TEE or CT within 24 h, uninterrupted OAC for ≥4 weeks, ACT ≥350 s, exclusion of CHA₂DS₂-VASc ≥4 — after which no strokes occurred. (sources/PFA-LPeAF-AVANT-GUARD-NEJM-2026, rating: very high)

CASTLE-HTx — ablation in end-stage HFrEF.

• The first RCT of catheter ablation in end-stage HFrEF (LVEF ≤35%, NYHA ≥II, referred for transplant or LVAD evaluation) with symptomatic AF; stopped early by the DSMB for overwhelming efficacy after a median 18 months. (sources/CA-HF-CASTLEHTx-NEJM-2023, rating: very high)

• Design. PVI alone in 53%, PVI plus additional substrate in 37%, multiple procedures in 9%; median 20 days from randomisation to ablation.
• Safety. Three vascular access complications versus one; no atrioesophageal fistulae and no tamponade.
• Limitations. Single-centre, open-label, stopped early at a median 18 months of a planned 3 years, with 16% crossover in the control arm — the effect size may be inflated and needs multicentre replication.
• Clinical implication. Ablation should be considered before transplant listing, not as an alternative to it, and should not postpone listing. (sources/CA-HF-CASTLEHTx-NEJM-2023)

Technique and energy sources

Two energy families are in use: non-thermal pulsed field ablation, which now carries most of the randomised evidence, and the thermal modalities — contact-force radiofrequency, cryoballoon and high-power short-duration RF.

Pulsed field ablation. Non-thermal electroporation: ultra-short high-voltage pulses open pores in myocardial membranes, killing the tissue while sparing the oesophagus, phrenic nerve and PV walls.

Full detail: concepts/PFA-Atrial-Arrhythmias.

• The two landmark RCTs. ADVENT (n=607, 30 centres, Bayesian adaptive) met non-inferiority for efficacy (73.3% versus 71.3%; posterior probability >0.999) and safety (2.1% versus 1.5%), and was superior for PV preservation (−0.9% versus −12.0% cross-sectional area); procedures were shorter but fluoroscopy longer. SINGLE SHOT CHAMPION (n=210, all with continuous monitoring) met non-inferiority and formal superiority over cryoballoon (37.1% versus 50.7% recurrence days 91–365; −13.6 pp; P=0.046, borderline), with lower recurrence even during the blanking period (−20.0 pp). (sources/pfa-advent-nejm-2023, rating: very high) (sources/pfa-cryo-singleshotchampion-nejm-2025)
• Reconciling the two trials. ADVENT operators were PFA-naïve, used intermittent Holter monitoring and compared against mixed RF and cryo; SINGLE SHOT CHAMPION operators were experienced, used continuous monitoring and compared against cryoablation alone.
• Aggregate evidence. Five RCTs now exist (ADVENT, SINGLE SHOT CHAMPION, BEAT PAROX-AF, INSIGHT PFA, SPHERE Per-AF): non-inferiority throughout, superiority only in SINGLE SHOT CHAMPION, shorter procedure times in every trial, and no difference in AF burden, quality of life or healthcare utilisation. PULSED AF and PersAFOne confirmed safety and efficacy in paroxysmal and persistent AF. (sources/pfa-af-ehra-2026, rating: very high)
• Safety. More than 500,000 procedures without a single reported oesophageal fistula (thermal incidence 0.025%, mortality 66%), no persistent phrenic nerve injury, no PV narrowing with PFA-only devices, and major adverse events of 0.9–1.9% across MANIFEST-PF, MANIFEST-17K and EU-PORIA. (sources/pfa-af-ehra-2026)
• PFA is not a class effect. Waveform, voltage, pulse duration and geometry differ more between PFA systems than between RF systems; there is no validated lesion-effect marker; dosing minima are manufacturer-defined without dose-response evidence; and no outcome data exist beyond 1 year. Recurrence during the blanking period was 38.1% with PFA versus 58.1% with cryoballoon, and early recurrence predicted late failure in PULSED AF, so a 1-month blanking period may be reasonable after PFA. (sources/pfa-af-ehra-2026)
• Limits beyond PVI. Durability outside the pulmonary veins is poor and no outcome data exist beyond 1 year, so PVI is where PFA is best established; the durability figures are under Lesion sets and targets. (sources/pfa-af-ehra-2026)

Energy comparison.

• Modality may affect progression. One multicentre RCT found PAF-to-persistent progression of 0% with CF-guided RF versus 7% (cryo, 4-minute applications) and 4.3% (cryo, 2-minute), P=0.03. (sources/ca-af-ehj-2024, rating: high)
• HPSD caveat. For left atrial posterior wall adjacent to the oesophagus, ≤5 s at high power is insufficient for durable lesions; oesophageal displacement strategies are required. (sources/ca-af-ehj-2024)

Transseptal access after septal closure

Left atrial access is the limiting step when the fossa ovalis is covered by an occluder or its anatomy is altered, and device dislodgement is the feared complication.

• Feasible in experienced hands. Pooled data from three single-arm series (64 patients, follow-up 6–22 months): transseptal puncture succeeded in 100%, 18.75% through the device under fluoroscopic plus intracardiac echocardiographic guidance, with 77.7% freedom from AF (95% CI 65.7–86.3), no device dislodgement and no residual shunt. Puncture through the device lengthened procedure time (237.3 versus 180 min, P=0.004) without changing recurrence. (sources/af-ablation-asd-device-jce-2020, rating: medium)
• The complication rate is the caution. Inadvertent epicardial puncture occurred in 4.7% (3 of 64) even in high-volume centres, and pooled observation supports safety only from 6 months after implantation — 64 patients across three uncontrolled series cannot establish a safety margin. Full detail: concepts/Atrial-Septal-Defect.

Procedural complications

• Major complications occur in about 5% overall: atrioesophageal fistula 0.2% and often fatal; cardiac tamponade 0.4–1.5%; stroke or TIA <1%; pulmonary vein stenosis 0.1–0.8%; death 0.1–0.4%. (sources/AF-AHA-2023, rating: very high)
• Energy-specific hazards are set out in the comparison table above — coronary vasospasm and haemolysis-related kidney injury with pulsed field; atrioesophageal fistula and steam pops with high-power RF; phrenic nerve palsy with cryoballoon. The distribution is not even: POTTER-AF found oesophageal fistula more prevalent with RF than cryoenergy. (sources/pfa-af-ehra-2026, rating: very high)

Lesion sets and targets

• Antral rather than ostial PVI. Antral isolation encompasses a wider area including part of the left atrial posterior wall; a meta-analysis of 1,183 patients found significantly better long-term arrhythmia freedom than ostial isolation. (sources/ca-af-ehj-2024, rating: high)
• Posterior wall isolation is unresolved. CAPLA failed to show benefit in ablation-naïve persistent AF, whereas a meta-analysis of 8 RCTs found reduced AF recurrence with LAPWI versus PVI plus roof line — discordance attributed to patient selection, non-transmural lesion failure and technique variation. Triggers below the inferior PV line cause recurrence if untargeted, and the septopulmonary bundle creates epicardial conduction gaps during roof-line ablation. (sources/ca-af-ehj-2024)
• Non-PV triggers. Common sites are the left atrial posterior wall, SVC, coronary sinus, crista terminalis, interatrial septum and LAA, more prevalent in non-paroxysmal AF, women, obesity, sleep apnoea, older age, low LVEF, severe LA scarring, HCM and mechanical mitral valves. Complete isolation is preferred to focal ablation at the SVC, LAA and CS. (sources/ca-af-ehj-2024)
• LAA isolation. BELIEF showed empirical LAA isolation improved long-term arrhythmia freedom in long-standing persistent AF, but it requires uninterrupted anticoagulation or an occlusion device. After multiple failed PVI, non-PV triggers are responsible in the majority, and adjunctive SVC, LAA or CS isolation restores high arrhythmia-free survival. In HCM, SVC and LAA isolation should be considered at redo if non-PV triggers are identified. (sources/ca-af-ehj-2024)
• Durability beyond the veins, with PFA. Posterior wall isolation achieved verified transmurality with no oesophageal injury but no 12-month benefit in MANIFEST-PF; mitral isthmus block persisted in only 5.5% at remapping with the pentaspline catheter; LAA isolation reconnected in 95% at 3 months; CTI ablation carries right coronary spasm and should be performed with IV nitroglycerin. (sources/pfa-af-ehra-2026)

Periprocedural anticoagulation

• During the procedure. Uninterrupted OAC is recommended (ESC Class I/A), or uninterrupted or minimally interrupted DOAC (AHA Class I/A). (sources/AF-ESC-2024, rating: very high)
• After the procedure. OAC continues for at least 3 months regardless of rhythm outcome. Long-term OAC is decided by thromboembolic risk — CHA₂DS₂-VA for ESC, CHA₂DS₂-VASc ≥2 for AHA — and is not stopped because ablation succeeded. (sources/AF-ESC-2024, sources/AF-AHA-2023)

OPTION — LAAO versus indefinite OAC after ablation.

• First large RCT of LAAO in patients undergoing catheter ablation with CHA₂DS₂-VASc ≥2 (men) or ≥3 (women): n=1,600, 106 sites, 10 countries, Watchman FLX versus OAC (95% NOAC). LAAO was concomitant with ablation in 40.9% and sequential at 90–180 days otherwise; post-implant OAC plus aspirin for 90 days, then aspirin alone to 12 months. (sources/laao-option-nejm-2025, rating: very high)

• The 18.1% bleeding rate in the OAC arm, despite a mean HAS-BLED of 1.2, shows that long-term DOAC therapy generates substantial clinically relevant bleeding even in low-risk post-ablation patients.
• LAAO at the time of ablation is therefore a viable way to remove lifelong anticoagulation while maintaining equivalent stroke protection, at a low procedural cost. (sources/laao-option-nejm-2025)

Surgical and hybrid ablation

• Persistent AF refractory to AADs: endoscopic or hybrid ablation, Class IIa/A with an experienced team and shared decision. (sources/AF-ESC-2024)
• Paroxysmal AF after failed percutaneous ablation: endoscopic or hybrid ablation, Class IIb/B. (sources/AF-ESC-2024)
• Concomitant surgical ablation: Class I/A with mitral valve surgery, Class IIa/B with non-mitral cardiac surgery. (sources/AF-ESC-2024)

Outcomes beyond arrhythmia control

• Quality of life. CABANA (n=2,204) showed significant improvement in AFEQT and Mayo AF symptom inventory scores at 12 months, sustained over time and independent of recurrence. STOP-AF, EARLY-AF and Cryo-FIRST showed quality-of-life gains with cryoablation; PFA similarly reduces AF burden with lower healthcare utilisation. (sources/ca-af-ehj-2024, rating: high)
• Cognition. A meta-analysis of 15,886 ablated versus 42,684 medically managed patients found HR 0.60 for incident dementia (95% CI 0.42–0.88; P<0.05). Acute MRI-detected cerebral lesions are frequent peri-procedurally but transient, with full cognitive recovery by 12 months. (sources/ca-af-ehj-2024)
• Progression. EARLY-AF (NEJM 2023) found cryoablation as first-line therapy reduced progression from paroxysmal to persistent AF from 7.4% with AADs to 1.9%. The mechanism is the "AF begets AF" paradigm: AF-induced structural remodelling drives progression, and early ablation interrupts the cycle. Energy choice may modify the effect further (above). (sources/ca-af-ehj-2024)

Genetic determinants of outcome

(Table: sources/genetic-af-dxmx-jce-2022, rating: medium)

• Proposed mechanism. 4q25 SNPs appear to increase LA scar formation and non-PV trigger frequency, reducing ablation effectiveness. (sources/genetic-af-dxmx-jce-2022)
• The evidence is retrospective. Every AF–ablation SNP association comes from retrospective observational studies; no prospective RCT has used genotype to guide ablation patient selection or technique, and the conflicting rs10033464 data show the risk of applying single-cohort findings clinically. (sources/genetic-af-dxmx-jce-2022)
• The largest study is negative. A 10-centre study of 3,259 patients with 1-year follow-up found no significant association between any common AF SNP, including 4q25, and ablation outcomes — the best-powered evidence, and it substantially weakens the "ablatogenomics" concept for common variants. (sources/genetic-af-cjc-2024, rating: high)
• LMNA cardiomyopathy. Pathogenic LMNA variants with cardiomyopathy carry dismal ablation outcomes, supporting a preference for rate control over rhythm control. (sources/genetic-af-cjc-2024)
• Family history of AF before age 65 in a first-degree relative predicts arrhythmia recurrence after ablation, independent of genotype-specific testing. (sources/genetic-af-dxmx-jce-2022)

Genotype-positive HCM (MYBPC3, MYH7).

• Gene-positive patients had more LA low-voltage area (12.3% versus 5.7%; normal voltage 87.7% versus 94.3%, P<0.001) despite comparable LA pressures and volumes, indicating primary genetic atrial myopathy rather than haemodynamic remodelling. (sources/MYBPC3-MYH7-JACCEP-2024, rating: medium)
• Freedom from AF at 12 months was comparable (75% versus 73%; P=0.92), but gene-positive patients needed more procedures (mean 1.67 versus 1.20; P=0.03) and all remained on AADs.
• At redo, 5 of 7 gene-positive patients showed interval progression of LA fibrosis and 71% had PV reconnections — attributed to non-transmural lesions in thicker LA walls and ongoing atrial myopathy.
• MYH7 variant carriers show a predilection for dense scar in the lateral LA; MYBPC3 carriers have greater LV mass.
• Non-PV trigger sources (posterior wall, SVC) are more prevalent in HCM, and SVC or LAA isolation should be considered at redo. (sources/ca-af-ehj-2024)
• See also concepts/HCM-Atrial-Myopathy for the mechanism.

Contradictions / Open Questions

• CABANA versus CASTLE-AF and CASTLE-HTx. CABANA showed no significant mortality benefit of ablation in its per-protocol analysis, while the HFrEF and end-stage HFrEF trials showed large benefit. Patient selection appears decisive: sicker, HF-predominant patients derive the most. (sources/AF-ESC-2024, sources/AF-AHA-2023, sources/CA-HF-CASTLEHTx-NEJM-2023)
• CASTLE-HTx effect size versus external validity. An HR of 0.24 for the primary composite is far larger than in any prior AF ablation trial and must be read cautiously — single-centre, early termination and 16% control crossover can all inflate benefit. Multicentre replication is needed. (sources/CA-HF-CASTLEHTx-NEJM-2023, rating: very high)
• Indefinite OAC versus LAAO after ablation. Both guidelines mandate indefinite OAC by stroke risk rather than ablation success, yet OPTION showed equivalent stroke protection with a 56% relative reduction in bleeding (HR 0.44) for LAAO at the time of ablation. Whether LAAO should formally replace the blanket anticoagulation mandate is unresolved, and so is whether OAC can ever be stopped after successful ablation. (sources/laao-option-nejm-2025, sources/AF-ESC-2024, rating: very high)
• Posterior wall isolation. CAPLA found no benefit in ablation-naïve persistent AF and MANIFEST-PF found none at 12 months with PFA, while a meta-analysis of 8 RCTs and RF/cryo registry data suggest benefit. Patient selection, transmurality and technique are the candidates, but the question is unreconciled. (sources/ca-af-ehj-2024, sources/pfa-af-ehra-2026)
• PFA superiority versus non-inferiority. SINGLE SHOT CHAMPION (n=210, continuous monitoring, experienced operators) showed borderline superiority over cryo (P=0.046); ADVENT (n=607, intermittent Holter, PFA-naïve operators) showed non-inferiority only. Whether the signal is real or a product of monitoring sensitivity and operator experience awaits a larger confirmatory trial. (sources/pfa-cryo-singleshotchampion-nejm-2025, sources/pfa-advent-nejm-2023)
• Blanking period after PFA. SINGLE SHOT CHAMPION found lower recurrence even during the 90-day blanking period with PFA (−20.0 pp versus cryo), suggesting less ablation-related inflammation; whether a blanking period is needed at all for PFA-based PVI is unresolved. (sources/pfa-cryo-singleshotchampion-nejm-2025, sources/pfa-af-ehra-2026)
• Persistent AF strategy. Large variability in technique persists, with no RCT defining optimal substrate modification or adjunctive lesion sets. (sources/AF-ESC-2024)
• Sham control. No sham-controlled ablation trials exist to separate a procedural placebo effect from true benefit on quality of life. (sources/AF-ESC-2024)
• RF versus cryo and AF progression. Whether modality choice meaningfully affects progression from paroxysmal to persistent AF rests on a single RCT and needs prospective validation. (sources/ca-af-ehj-2024)
• ESC versus AHA Class I scope. ESC Class I/A covers first-line ablation in paroxysmal AF broadly; AHA restricts Class I/A to "selected younger patients with few comorbidities" and assigns IIa/B-R to the broader paroxysmal and persistent population. (sources/AF-AHA-2023, sources/AF-ESC-2024)
• 4q25 and ablation outcomes. The association is hypothesis-generating only: all data are retrospective, rs10033464 conflicts across populations, and the largest study (n=3,259) found no association between any common AF SNP and ablation outcomes. (sources/genetic-af-dxmx-jce-2022, sources/genetic-af-cjc-2024, rating: medium and high respectively)

Connections

• concepts/PFA-Atrial-Arrhythmias — the energy platform in detail: mechanism, systems, dosing, workflow, trial evidence and safety; this page keeps only the AF-facing summary
• concepts/Ventricular-PFA — the ventricular application of the same energy
• entities/LAAO — OPTION: LAAO at the time of ablation as an alternative to indefinite OAC
• concepts/CHA2DS2-VA — the stroke-risk score that governs long-term anticoagulation after ablation
• concepts/Tachycardia-Induced-Cardiomyopathy — the HFrEF subgroup in which ablation is curative
• concepts/AF-CARE — the guideline framework surrounding the rhythm-control pathway
• concepts/AF-Staging — stage of disease influences when ablation is offered
• concepts/AF-Burden — the endpoint used in AVANT GUARD and CASTLE-HTx
• concepts/Antiarrhythmic-Drugs — the comparator in first-line trials
• concepts/Genetic-Testing-in-AF — 4q25 and other variants in AF risk and treatment response
• concepts/HCM-Atrial-Myopathy — the fibrotic substrate in sarcomeric HCM
• entities/Atrial-Fibrillation — the parent disease page
• entities/HCM, entities/MYBPC3, entities/MYH7 — the genotype-positive HCM evidence

Sources

• sources/pfa-af-ehra-2026
• sources/PFA-LPeAF-AVANT-GUARD-NEJM-2026
• sources/pfa-advent-nejm-2023
• sources/pfa-cryo-singleshotchampion-nejm-2025
• sources/CA-HF-CASTLEHTx-NEJM-2023
• sources/laao-option-nejm-2025
• sources/AF-AHA-2023
• sources/AF-ESC-2024
• sources/af-ablation-asd-device-jce-2020
• sources/EAST-AFNET4-NEJM-2020
• sources/MYBPC3-MYH7-JACCEP-2024
• sources/ca-af-ehj-2024
• sources/genetic-af-cjc-2024
• sources/genetic-af-dxmx-jce-2022