Safety of pulsed field ablation in more than 17,000 patients with atrial fibrillation in the MANIFEST-17K study

Atrial fibrillation (AF) is the most common sustained arrhythmia with substantial impacts on quality of life, morbidity, and mortality. Catheter ablation using thermal energy (radiofrequency or cryotherapy) is effective but intrinsically non-selective, risking collateral injury to adjacent structures such as the esophagus, pulmonary veins, and phrenic nerve, and rare but catastrophic atrio-esophageal fistula. Pulsed field ablation (PFA) delivers ultrarapid electrical pulses to induce irreversible electroporation, with pre-clinical and early clinical data suggesting relative selectivity for myocardial tissue and sparing of adjacent structures. The MANIFEST-17K study aimed to evaluate the real-world safety of post-approval PFA using a pentaspline catheter across a large, multinational cohort, testing whether PFA’s tissue preferentiality translates into reduced complications at scale.

Early human trials (IMPULSE/PEFCAT/PEFCAT2 and PersAFOne) established feasibility and initial safety of PFA in small cohorts (<150 patients). Post-approval registries (MANIFEST-PF, n=1,758; EU-PORIA, n=1,233) reported no esophageal injury or pulmonary vein stenosis and rare phrenic nerve palsy. The ADVENT randomized trial (n=707 paroxysmal AF) showed noninferiority of PFA versus thermal ablation for safety and efficacy. Historical experience with cryoballoon ablation showed that rare complications (e.g., atrio-esophageal fistula) may emerge only after thousands of cases, underscoring the need for large-scale surveillance. Pre-clinical models demonstrated esophageal and phrenic nerve sparing with PFA compared with radiofrequency ablation. Comparative imaging and clinical studies have shown less PV narrowing with PFA than with RF, and less esophageal injury on MRI. Reports have described coronary vasospasm during ablation near coronary arteries across energy sources, with prophylactic nitroglycerin showing benefit.

Design: Retrospective observational, center-level registry (MANIFEST-17K) of post-approval PFA with a pentaspline catheter (Farawave, Farapulse–Boston Scientific). Ethics approval: Homolka Hospital ethical committee (6.4.2022/18); informed consent waived due to retrospective design. Centers and patients: Invitations sent to 116 centers; 106 participated (91.4%). Data covered procedures mainly between early 2022 and June 2023. Initial 1,758 patients from 2021 reported in MANIFEST-PF were excluded. Final cohort: 17,642 patients (mean age 64 years; 34.7% female; AF type: paroxysmal 57.8%, persistent 35.2%, long-standing persistent 5.6%; atrial flutter/AT 1.4%). Procedures performed under general anesthesia in 46.9% and deep sedation without intubation in 53.1%. Data collection: Institution-level forms captured site characteristics, limited baseline and procedural parameters, and all adverse events (AEs). For specific AEs (stroke, coronary spasm), root cause/AE detail forms were obtained. No esophageal management strategies were used during PFA per protocol. PFA procedure: A 12F over-the-wire pentaspline catheter advanced via a 13F steerable sheath into the LA. Guidewire positioned in each PV; eight pulses per vein (four in basket and four in flower configurations with rotations). Posterior wall ablation delivered with overlapping flower poses. Typical PF voltage 2.0 kV (range 1.8–2.0 kV). AE definitions: Major AEs included death, stroke, esophageal fistula/dysmotility, PV stenosis, persistent phrenic nerve injury, pericardial tamponade, vascular complications requiring intervention, coronary spasm, myocardial infarction, hemolysis-related renal failure requiring hospitalization/dialysis, and thrombosis. Minor AEs included transient ischemic attack (TIA), transient phrenic nerve injury, pericardial effusion not requiring intervention, non-intervention vascular complications, hemolysis without hospitalization, and other events. Statistics: Descriptive analyses; comparisons by t-test or Mann–Whitney U for continuous variables and chi-squared/Fisher exact for categorical variables. Two-tailed P<0.05 considered significant. Analyses performed with SPSS v29.

- Cohort and participation: 106/116 centers (91.4%) contributed data on 17,642 patients (mean age 64; 34.7% female; paroxysmal AF 57.8%, persistent AF 35.2%). Average follow-up 15 months. - Major complications: 0.98% (173/17,642). - Pericardial tamponade: 0.36% (63/17,642); 88.9% managed percutaneously (56/63), 11.1% surgically (7/63). - Vascular complications requiring intervention: 0.30% (53/17,642). - Stroke: 0.12% (22/17,642); none fatal. In 16 cases with analysis, putative causes included catheter exchanges/sheath management (n=4), ACT<300 (n=2), anticoagulation interruption (n=2), uncontrolled hypertension (n=1), no definitive cause (n=7). Asymptomatic MRI lesions detected in 9/96 (9.4%) in a subset with routine MRI. - Coronary arterial spasm: 0.14% (25/17,642); 88% proximity-related (mitral isthmus or CTI), 12% generalized; EKG changes in 92%; hypotension 20%; IV/intracoronary nitroglycerin used in 84%; clinical sequelae in 16% including 2 ventricular fibrillation cases with recovery. - Death: 0.03% (5/17,642); two procedure-related (tamponade leading to neurological injury; post-procedure heart failure), three occurred 3–30 days post-ablation (two sudden deaths—one with severe CAD; one non-cardiovascular cerebellar hemorrhage). - Hemolysis-related acute renal failure requiring hemodialysis: 0.03% (5/17,642), all with complex lesion sets (PVI plus posterior wall and linear lesions) and high application numbers (~143±27 per procedure); all recovered with transient dialysis. - Minor complications: 3.21% (567/17,642): vascular (no intervention) 2.20% (388/17,642), pericardial effusion (no intervention) 0.33% (59/17,642), pericarditis 0.17% (30/17,642), TIA 0.12% (21/17,642), transient phrenic injury 0.06% (11/17,642; all resolved by discharge). - PFA energy-specific safety: 0 esophageal complications (no fistula/dysmotility), 0 pulmonary vein stenosis, 0 persistent phrenic nerve injury. - Ultrasound-guided access: Sites not routinely using ultrasound had higher major vascular complication rates (0.50% vs 0.17%, P=0.046), despite similar case volumes. - Learning curve: Compared with earlier MANIFEST-PF experience at the initial 24 sites, MANIFEST-17K showed numerically lower rates of major AEs (1.65% to 0.98%), tamponade, stroke, and transient phrenic paresis, with dissemination of best practices to new sites (community-level learning). - Comparative context: Rates for tamponade (0.36%), stroke (0.12%), and mortality (0.03%) are markedly lower than historical thermal ablation registries (tamponade ~1.52%, stroke ~1.02%, mortality ~0.42%).

In this large, multicenter real-world cohort, PFA with a pentaspline catheter demonstrated a favorable safety profile consistent with selective myocardial ablation and reduced collateral injury. The absence of esophageal complications, PV stenosis, and persistent phrenic nerve injury across >17,000 patients supports the tissue preferentiality suggested by pre-clinical and early clinical data. Non–energy-specific complications—principally tamponade, vascular events, and stroke—occurred at low rates and compared favorably with historical thermal ablation data, despite diverse operator experience and early adoption. Two PFA-related phenomena warrant specific attention. First, coronary vasospasm (0.14%) occurred mainly when ablating near coronary arteries (mitral isthmus or CTI), occasionally with serious sequelae. Prophylactic nitroglycerin may mitigate risk during proximity ablation. Second, hemolysis-associated acute kidney injury requiring temporary dialysis (0.03%) was observed in cases with extensive lesion sets and high numbers of applications, suggesting a dose-dependent effect related to electroporation of erythrocytes; preventive strategies (moderating application number, ensuring hydration) may reduce risk. A learning curve was evident at both center and community levels, with lower adverse event rates over time and in newly adopting centers, likely reflecting shared best practices (e.g., careful catheter manipulation, J-tip guidewire use, vigilant sheath management and flushing). Collectively, these findings address the central question of PFA safety at scale and indicate that the technology avoids much of the collateral damage associated with thermal ablation while maintaining low complication rates in routine practice.

The MANIFEST-17K registry, encompassing 17,642 consecutive real-world PFA cases across 106 centers, demonstrates that pentaspline PFA has a favorable safety profile with no esophageal complications, PV stenosis, or persistent phrenic nerve injury, and low rates of major complications (0.98%), stroke (0.12%), and mortality (0.03%). Rare but important findings include coronary vasospasm (mostly proximity-related) and hemolysis-related acute renal failure after extensive lesion sets. The observed learning curve and dissemination of best practices further support the safe adoption of PFA. Future work should refine procedural strategies to minimize vasospasm (e.g., nitroglycerin prophylaxis during proximity ablation), characterize and mitigate hemolysis risk (optimize application number, hydration), and continue surveillance to detect rare or subclinical events.

- Retrospective observational design with center-level data and without prospectively predefined endpoints may introduce reporting bias. - Potential underdetection of subclinical events (e.g., esophageal lesions, asymptomatic PV stenosis, silent cerebral ischemia, subclinical coronary spasm, mild hemolysis-related renal dysfunction) due to lack of routine post-procedure imaging/screening. - Diagnosis of vasospasm sometimes inferred clinically (prompt nitroglycerin, EKG changes, proximity), not always angiographically confirmed. - Operator cohort may be enriched for expertise; safety outcomes may not generalize to all centers. - Findings pertain to the pentaspline PFA catheter/system; results should not be extrapolated to other PFA technologies without validation.