A systematic analysis of 2,262 adverse event reports from the FDA's MAUDE database — 760 involving pulsed field ablation (PFA) and 1,502 involving radiofrequency (RF) ablation — found that PFA's safety profile is not monolithic. Pooled PFA showed meaningfully lower reporting odds for cardiac tamponade (ROR 0.52) and esophageal injury (ROR 0.09) compared to RF, confirming its tissue-selective advantage. However, Varipulse carried a dramatically elevated stroke signal (ROR 16.41, 95% CI 8.61–31.28) that was entirely absent for Farapulse (ROR 1.26, non-significant). PFA-exclusive complications included coronary vasospasm (24 vs. 0 events) and hemolysis (15 vs. 1 events).
This preprint, not yet peer-reviewed, challenges the clinical assumption that FDA approval of a device class implies interchangeable safety. For the estimated 200,000+ AF ablation procedures performed annually in the U.S., platform selection now carries concrete, data-supported risk implications — not merely technical preference. The stroke signal attenuation seen after an FDA Safety Communication (ROR falling to 1.60, non-significant) illustrates how notoriety bias can transiently distort pharmacovigilance databases, a well-documented limitation of MAUDE analyses. Importantly, MAUDE is passive, voluntary, and subject to underreporting, making causal inference impossible. Findings should be validated against prospective registries. Still, as an early signal-generation tool, this analysis is practically significant: electrophysiologists and patients choosing between approved PFA platforms now have real-world differentiation data, an incremental but clinically meaningful advance.