Reclassifying Ablation Energy Risk Profiles in Atrial Fibrillation: A Narrative-Adjudicated Comparative Analysis of the MAUDE Database
- 1The University of Utah School of Medicine
Abstract
BackgroundPulsed field ablation (PFA) has been rapidly adopted for atrial fibrillation (AF) on the premise of greater safety than thermal ablation. Real-world signals of hemolysis, coronary spasm, and neurologic events have since emerged, and prior Manufacturer and User Facility Device Experience (MAUDE) analyses relied on term matching without prospective signal detection.
MethodsWe queried MAUDE for AF ablation catheter reports received January 2016 through June 2026 and assigned each to PFA, radiofrequency ablation (RFA), or cryoballoon ablation. We adjudicated injury and death narratives into predefined complication families and severity tiers with a large language model, checked against a larger model and by physician review. We calculated reporting odds ratios (ROR) and Bayesian shrinkage estimators compared against an empirical null, applied interrupted time series and prospective sequential detection, and repeated estimates on a 2026 hold-out.
ResultsAmong 44,623 reports, 30.1% were PFA, 50.9% RFA, and 19.0% cryoballoon, with 91.8% model agreement on event occurrence ({kappa} = 0.83). PFA carried an ROR of 527 (95% CI, 33, 8,509) for hemolysis, 2.83 (2.55, 3.16) for arrhythmia, 2.83 (2.30, 3.48) for coronary, and 2.39 (2.10, 2.73) for cerebrovascular events, against 0.024 (0.008, 0.075) for esophageal and 0.34 (0.32, 0.37) for pericardial events. The cerebrovascular excess was largest for VARIPULSE, but without it the class proportional reporting ratio was 1.36 (1.13, 1.65), concordant with the 1.56 measured in EMBOL-AF. In the hold-out, the cerebrovascular, arrhythmia and hemolysis signals replicated, and coronary events kept their direction without meeting the signal rule. Prospective detection flagged cerebrovascular events 13 months before AVANT GUARD published its neurologic events.
ConclusionsPFA showed disproportionate reporting of hemolysis, cerebrovascular, coronary, and arrhythmic events across sequential testing. Machine learning methods allowed for flagging these signals months before confirmatory trials, demonstrating a scalable tool for early detection of procedural risk.
CLINICAL PERSPECTIVEWhat is Known O_LIPulsed field ablation has been adopted rapidly for atrial fibrillation on the premise of fewer collateral injuries than thermal ablation. C_LIO_LIHemolysis, coronary spasm, and neurologic events have since emerged in case series, registries, and a paused trial, while prior analyses of Food and Drug Administration device reports relied largely on term matching without narrative adjudication at scale. C_LI What the Study Adds O_LIIn injury and death reports adjudicated at the narrative level, pulsed field ablation was reported more often than thermal ablation for hemolysis, arrhythmia, coronary, and cerebrovascular events and less often for esophageal and pericardial injury. C_LIO_LIThe cerebrovascular reporting excess was largest for the VARIPULSE catheter but persisted with that platform removed and concentrated in minor stroke and systemic embolism, suggesting an embolic signal not confined to one platform. C_LIO_LISequential detection run on reports adjudicated by a large language model would have flagged the cerebrovascular signal 13 months before AVANT GUARD published its neurologic events, suggesting a role for automated surveillance of new ablation technologies. C_LI
Related
- Comparative efficacy of esophageal protection strategies during radiofrequency atrial fibrillation ablation: a systematic review and network meta-analysis of randomized trials.
- Reclassifying No Specific Molecular Profile Endometrial Cancer: A Narrative Review of Molecular Heterogeneity, Prognostic Refinement, and Clinical Management
- Pulsed Field Versus Radiofrequency Ablation For Repeat Ablation of Atrial Fibrillation: Evaluation of Outcomes
- Clinical validation of a novel catheter–tissue contact algorithm in atrial fibrillation ablation using pulsed field ablation
- Atrial fibrillation phenotypes derived from multimodal data exhibit distinct electrophysiological substrate characteristics, symptom trajectories, and ablation outcomes
BibTeX
@misc{shah2026reclassifying,
title = {{Reclassifying Ablation Energy Risk Profiles in Atrial Fibrillation: A Narrative-Adjudicated Comparative Analysis of the MAUDE Database}},
author = {A. S. Shah and R. Fuller and E. G. Daoud and B. Endo and A. Goodwin and R. Martinez-Parachini and M. Engels and H. I. Lehmann and K. Konstantinidis and R. Ranjan and T. J. Bunch},
howpublished = {medRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.30.26364455},
url = {https://www.medrxiv.org/content/10.64898/2026.09.30.26364455v1},
note = {CC BY-ND},
}
RIS
TY - UNPB
TI - Reclassifying Ablation Energy Risk Profiles in Atrial Fibrillation: A Narrative-Adjudicated Comparative Analysis of the MAUDE Database
AU - A. S. Shah
AU - R. Fuller
AU - E. G. Daoud
AU - B. Endo
AU - A. Goodwin
AU - R. Martinez-Parachini
AU - M. Engels
AU - H. I. Lehmann
AU - K. Konstantinidis
AU - R. Ranjan
AU - T. J. Bunch
PY - 2026
DA - 2026/09/30
PB - medRxiv
DO - 10.64898/2026.09.30.26364455
UR - https://www.medrxiv.org/content/10.64898/2026.09.30.26364455v1
AB - BackgroundPulsed field ablation (PFA) has been rapidly adopted for atrial fibrillation (AF) on the premise of greater safety than thermal ablation. Real-world signals of hemolysis, coronary spasm, and neurologic events have since emerged, and prior Manufacturer and User Facility Device Experience (MAUDE) analyses relied on term matching without prospective signal detection. MethodsWe queried MAUDE for AF ablation catheter reports received January 2016 through June 2026 and assigned each to PFA, radiofrequency ablation (RFA), or cryoballoon ablation. We adjudicated injury and death narratives into predefined complication families and severity tiers with a large language model, checked against a larger model and by physician review. We calculated reporting odds ratios (ROR) and Bayesian shrinkage estimators compared against an empirical null, applied interrupted time series and prospective sequential detection, and repeated estimates on a 2026 hold-out. ResultsAmong 44,623 reports, 30.1% were PFA, 50.9% RFA, and 19.0% cryoballoon, with 91.8% model agreement on event occurrence ({kappa} = 0.83). PFA carried an ROR of 527 (95% CI, 33, 8,509) for hemolysis, 2.83 (2.55, 3.16) for arrhythmia, 2.83 (2.30, 3.48) for coronary, and 2.39 (2.10, 2.73) for cerebrovascular events, against 0.024 (0.008, 0.075) for esophageal and 0.34 (0.32, 0.37) for pericardial events. The cerebrovascular excess was largest for VARIPULSE, but without it the class proportional reporting ratio was 1.36 (1.13, 1.65), concordant with the 1.56 measured in EMBOL-AF. In the hold-out, the cerebrovascular, arrhythmia and hemolysis signals replicated, and coronary events kept their direction without meeting the signal rule. Prospective detection flagged cerebrovascular events 13 months before AVANT GUARD published its neurologic events. ConclusionsPFA showed disproportionate reporting of hemolysis, cerebrovascular, coronary, and arrhythmic events across sequential testing. Machine learning methods allowed for flagging these signals months before confirmatory trials, demonstrating a scalable tool for early detection of procedural risk. CLINICAL PERSPECTIVEWhat is Known O_LIPulsed field ablation has been adopted rapidly for atrial fibrillation on the premise of fewer collateral injuries than thermal ablation. C_LIO_LIHemolysis, coronary spasm, and neurologic events have since emerged in case series, registries, and a paused trial, while prior analyses of Food and Drug Administration device reports relied largely on term matching without narrative adjudication at scale. C_LI What the Study Adds O_LIIn injury and death reports adjudicated at the narrative level, pulsed field ablation was reported more often than thermal ablation for hemolysis, arrhythmia, coronary, and cerebrovascular events and less often for esophageal and pericardial injury. C_LIO_LIThe cerebrovascular reporting excess was largest for the VARIPULSE catheter but persisted with that platform removed and concentrated in minor stroke and systemic embolism, suggesting an embolic signal not confined to one platform. C_LIO_LISequential detection run on reports adjudicated by a large language model would have flagged the cerebrovascular signal 13 months before AVANT GUARD published its neurologic events, suggesting a role for automated surveillance of new ablation technologies. C_LI
ER -