A single-centre study has found a significantly higher incidence of stroke and transient ischaemic attacks (TIA) within 30 days among patients who underwent pulsed field ablation (PFA) relative to those who received radiofrequency ablation (RFA).
“In the largest comparative safety analysis to date, PFA was associated with a fivefold higher stroke rate vs RFA,” said lead study author Dr Enrico G Ferro, Beth Israel Deaconess Medical Center, Boston, US.
“The PFA stroke signal was not explained by patient- or operator-level differences, nor by learning curve effects,” he added. [Ferro EG, et al, EHRA 2026]
Twelve operators performed PFA and RFA in 1,504 and 2,062 patients, respectively, between January 2022 and July 2025. The mean age of individuals in both arms was 67 years, and 31 percent were female. Comorbidity profiles did not significantly differ between the two groups: 45 percent paroxysmal and 50 percent persistent atrial fibrillation (AF); mean CHAâ‚‚DSâ‚‚-VASc 2.4, with 7 percent prior stroke/TIA; mean HAS-BLED 1.5.
Patients who received PFA vs RFA did so earlier in their disease course, as evidenced by the lower use of antiarrhythmic medications (75 percent vs 81 percent; p<0.001) and cardioversions (40 percent vs 32 percent; p<0.001).
Left atrial dwell time was much shorter with PFA than with RFA (1.8 vs 2.4 h; p<0.001). Pulmonary vein isolation (PVI)-only was conducted in 18.8 percent of PFA patients and 46 percent of RFA patients (p<0.001), whereas PVI/posterior wall isolation was performed in 57 percent and 25.5 percent, respectively (p<0.001), with similar quantities of additional lesions in the two arms (eg, mitral lines in approximately 20 percent of patients).
Stroke, ischaemic attacks
At 30 days, stroke/TIA events occurred in nine patients (0.20 percent), which were all confirmed by MRI or fundoscopy: 7 (0.47 percent) after PFA and 2 (0.10 percent) after RFA (unadjusted risk difference, 0.37 percent, 95 percent confidence interval [CI], 0.00–0.74; p=0.05).
The two arms remained well balanced following inverse probability of treatment weighting (IPTW), and PFA still showed a significant correlation with higher rates of stroke and TIA at both days 7 (0.43 percent vs 0.05 percent; 95 percent CI, 0.04–0.72; p=0.03) and 30 (0.51 percent vs 0.10 percent; 95 percent CI, 0.02–0.79; p=0.03).
In terms of safety, atrioesophageal fistulas were not observed, and the adjusted rates of other complications did not significantly differ between the PFA and RFA procedures: phrenic nerve injury (0 percent vs 0.15 percent; p=0.34), coronary vasospasm (0 percent vs 0.05 percent; p=0.60), and pericardial tamponade (0.07 percent vs 0.24 percent; p=0.34).
“While mechanistic definitions are not definitive, the current PFA strategy is associated with higher stroke risk,” Ferro said.
“Further studies are needed to confirm these findings, define mechanisms, and optimize procedural protocols to mitigate the cerebrovascular risk,” he added.
Ferro and his team conducted this study by analysing a prospective registry from a high-volume US academic centre, including all consecutive AF ablations from 2022 to 2025. They recorded all complications arising from the procedures. Three independent neurologists blinded to treatment assignment adjudicatedstroke and TIA events.
Ferro and colleagues applied IPTW to balance clinical differences between the PFA and RFA arms. They also compared stroke/TIA rates at 30 days (primary endpoint) using propensity scores. Other endpoints assessed were nonfatal complications at 30 days and stroke/TIA rates at 7 days.
The study was limited by its single-centre and nonrandomized design. Moreover, the low event rates limited the power of the mediation analyses, according to Ferro.