Interventional procedure overview of pulsed field ablation for atrial fibrillation

Population and studies description

The identified publications were stratified by technology and then prioritised to select the key evidence that is representative of this procedure. Evidence was prioritised for 5 technologies. Key evidence for other technologies is presented in the first section of appendix B, table 5.

This interventional procedures overview is based on 12 studies reported in 17 publications. They are presented in table 2 and table 3, and 100 other relevant studies are presented in appendix B, table 5. The publications include about 30,000 people of whom about 24,850 had PFA, from the UK, EU, Israel, US, Canada, Australia and China. Most studies reported outcomes up to 12 months. One study reported acute outcomes within 48 hours as the focus was on prevention of a common periprocedural adverse event (Mohanty 2024). One study of 17,642 people had average follow up of 15 months and data up to 25 months (Ekanem 2024). The UK analysis of 707 people comparing PFA with RFA and CBA (Calvert 2024) reported findings at a median of 7.6 months.

The selected evidence included 2 systematic review and meta-analyses (de Campos 2024, Amin 2024), mostly comprised of evidence from the Farawave catheter (Boston Scientific, see Procedure technique for more information on the different technologies in the prioritised evidence). De Campos (2024) included 18, mostly non-randomised comparative studies with 4,998 people (1,761 had PFA), and Amin (2024) included 17 mostly non-randomised comparative studies with 2,255 people (1,051 had PFA). Both publications compared PFA with TA (RFA and CBA) and both included evidence from across the globe. Eight studies were included in both systematic review and meta-analyses. The differences in included studies was likely due to de Campos (2024) having a more recent search date and because they only included studies done in people having first-time ablation. Also, Amin (2024) did not have an exclusion criteria around repeat ablation procedures.

Other prioritised evidence done with the Farawave catheter was from a US 3-arm multicentre RCT comparing PFA with RFA and CBA (ADVENT trial, Reddy 2023a, Patel 2024), a UK single-arm real-world study of 707 people comparing RFA with PFA and CBA (Calvert, 2024), an EU retrospective analysis of a prospective registry of 1,568 people across 24 centres (MANIFEST-PF study, Turagam 2023, Turagam 2024), a centre-level study done across 101 centres in the EU and Israel including 17,642 people (MANIFEST-17K study, Ekanem 2024), a prospective case series of 191 people who all had PFA with deep sedation (Schmidt 2022), and a retrospective case series of 103 people focusing on avoidance of a common adverse event (Mohanty 2024).

Prioritised evidence from the Varipulse catheter (Biosense Webster Inc.) was from an EU and Canada prospective single-arm trial of 226 people across 13 centres (inspIRE study, Duytschaever 2023, de Potter 2024, Grimaldi 2023).

Prioritised evidence for the PulseSelect catheter (Medtronic) was from an international prospective single-arm trial of 300 people across 41 centres (PULSED AF, Verma 2023a, Verma 2023b).

Prioritised evidence for the Sphere-9 catheter (Affera Inc, Medtronic) was from a publication that combined evidence from 2 prospective single-arm trials which together had 178 people across 3 EU centres (Reddy 2023b).

Prioritised evidence for the CENTAURI PFA generator (Galvanize therapeutics) was from a prospective, single-arm trial of 84 people from 2 EU centres (Anić 2023).

Most studies included a mixture of people who had PAF, PersAF and some also reported including people with longstanding PersAF. Some studies did separate analyses for these subgroups. The ADVENT RCT comparing PFA with RFA and CBA (Reddy 2023a, Patel 2024) and the inspIRE single arm trial of 226 people (Duytschaever 2023, de Potter 2024, Grimaldi 2023), only included people with PAF.

The evidence included a mixture of people having first-time PFA and people having repeat procedures. The systematic review and meta-analysis of 18 studies comparing PFA with TA only included people having first-time PVI ablation (de Campos 2024). The systematic review and meta-analysis of 17 studies comparing PFA with TA included studies of people having both first-time and repeat ablation procedures but the proportion of each was not reported (Amin 2024). Most of the remaining studies only included people having first-time procedures. The EU-PORIA centre-level registry study did not exclude people that had had previous ablation procedures (Schmidt 2023), but most people (96%) had first time ablation, and inclusion criteria with respect to repeat procedures was unclear in Mohanty (2024). Most studies reported need for repeat ablation procedure after the index PFA procedure as an outcome.

This is a rapid review of the literature, and a flow chart of the complete selection process is shown in figure 1. Table 2 presents study details.

Procedure technique

Each PFA technology has unique design and properties, briefly described below. Some studies only used GA, whilst others used a mixture of GA and conscious or deep sedation. Schmidt (2022) and the inspIRE sub study (Grimaldi 2023) both used deep sedation only. Some studies only applied PFA for PVI whilst others included people who had PFA for other lesion sets (like left atrial posterior wall ablation), at the operator's discretion. Typically, uninterrupted oral anticoagulation was used. Use of 3D mapping systems was used variably within and between studies, whilst some procedures were done under fluoroscopic guidance only. One study explicitly reported differences in haemoglobinuria and acute kidney injury outcomes when post-procedure hydration was or was not used (Mohanty 2024).

Farawave
The Farawave catheter was used in 6 studies. These were the ADVENT RCT (Reddy 2023a, Patel 2024), the UK real-world study of 707 people comparing RFA with PFA and CBA (Calvert, 2024), the MANIFEST-PF EU retrospective analysis of a prospective registry of 1,568 people across 24 centres (Turagam 2023, Turagam 2024), the MANIFEST-17K centre-level study of 101 centres with 17,642 people (Ekanem 2024), the prospective case series of 191 people (Schmidt 2022) and retrospective case series of 103 people (Mohanty 2024).

This catheter is referred to as a 'single shot' technology. It is available in 2 sizes, 31 mm or 35 mm diameter. It delivers a bipolar, biphasic waveform. A standard procedure protocol has been developed for this technology. The PFA catheter is used in a basket configuration to deliver 2 sets of PFA lesions, then rotated to deliver two additional PFA lesions, then changed to a flower configuration to deliver two more PFA lesions, and finally rotated again to deliver two final PFA lesions, for a minimum total of 8 PFA applications per PV. Each application is 5 pulse trains, delivered over 2.5 seconds. The lesions may be delivered at 1800, 1900 or 2000 volts.

In the ADVENT RCT (Reddy 2023a, Patel 2024), GA or deep sedation was used at the operator's discretion. In the UK real-world study of 707 people (Calvert, 2024) and the retrospective case series of 103 people (Mohanty 2024), GA only was used. In MANIFEST-PF(Turagam 2023, Turagam 2024), the authors described that intravenous moderate sedation or GA was used. In MANIFEST-17K (Ekanem 2024), procedures were done with deep sedation without intubation in 53% of people but the authors did not report what was done for the rest of the group. In prospective case series of 191 people (Schmidt 2022), all people had deep sedation.

Varipulse
In inspIRE (Duytschaever 2023, de Potter 2024, Grimaldi 2023), the Varipulse catheter was used. This catheter is referred to as a 'single shot' technology. The authors describe that this catheter has a bidirectional circular tip that can be expanded and contracted to fit PVs of different sizes. PFA with Varipulse is applied in a bipolar configuration with an energy of 1800 V and each application has microsecond-long square-wave biphasic pulses, for a total application duration of about 250 milliseconds. In this study, procedures were done under conscious sedation, deep sedation or GA. Deep sedation outcomes were reported in Grimaldi (2023). A mapping system was used and 12 applications of PFA were done per PV.

PulseSelect
In PULSED AF (Verma 2023a, Verma 2023b), the PulseSelect catheter was used. This catheter is referred to as a 'single shot' catheter. Each application was 4 biphasic, bipolar pulse trains, each lasting 100 to 200 milliseconds at 1400 to 1500 V measured from baseline to peak (2800 to 3000 V measured peak to peak). After each application, the catheter was rotated circumferentially to a new position to achieve full circumferential isolation. No system or waveform modifications were made during the trial. In this study, all procedures were done under conscious sedation or GA. Mapping systems were only used with operator discretion.

Sphere-9
In the publication that combined evidence from 2 prospective single-arm trials which together had 178 people across 3 EU centres (Reddy 2023b), the Sphere-9 catheter was used. This catheter is referred to as a 'point-by-point' catheter. Unlike other catheters, the system has a dual generator design so operators can toggle between RFA and PFA. During PFA applications, a biphasic monopolar PF waveform is delivered from the entire lattice tip. Waveforms consist of a train of microsecond-scale pulses delivered over 3 to 5.5 seconds, driven with up to plus or minus 2 kV. The length of lesion application time developed over time in this study. All procedures were done under GA. A mapping system was used in all procedures. PFA was used for all posterior left atrial applications lesion sets, but the anterior applications were done using either RFA or PFA. Outcomes were reported separately.

CENTAURI
In ECLIPSE-AF (Anic 2023) the CENTAURI generator was used. Unlike the other technologies, CENTAURI is a generator of PFA energy and it can be used with commercially available catheters and mapping systems, like those used for RFA. The CENTAURI generator delivers biphasic, monopolar pulsed electric field energy at three selectable energy settings (19, 22, and 25 A). During the study, workflows were optimised in phases, and with different catheters and mapping systems. Mapping systems were used in this study. The following catheters were used with the CENTAURI generator sequentially as the workflows developed: TactiCath SE, INTELLANAV STABLEPOINT, and THERMOCOOL SMARTTOUCH. Standard wide antral circumferential ablation was done with sufficient overlap of each lesion. A mixture of GA and deep sedation were used in this study.

Procedure details in the systematic review and meta-analyses
In the meta-analysis comparing PFA with thermal ablation (de Campos 2024) 17 of 18 studies used the Farawave catheter and 1 study used the 8F disposable cardiac PFA catheter (PFA8D18L). In the systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024), Farawave was used in all studies. No further procedure details were reported in either publication.

Efficacy

Composite scores of treatment success or failure
Four studies reported a composite score of treatment success or failure. These represented conservative estimates by combining status or response on several desirable clinical outcomes. Point estimates for treatment success were reported in 3 studies. Rates were between 71 and 73% in 2 studies but as low as 55% in the other, and conclusions about comparative effectiveness between PFA and TA differed between comparative studies. In 2 studies, findings were more favourable in people with PAF than PersAF but this was not corroborated by the systematic review findings. The differences in conclusions are likely primarily driven by differences in definitions of treatment success or failure.

In the systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024), data from studies were combined to measure overall treatment failure. The authors defined this as rate of treatment failure after 3 months of blanking period and by any reason reported, such as any atrial tachyarrhythmia recurrence, including AF, atrial tachycardia, and atrial flutter, repeat ablation, use of antiarrhythmic drugs, and cardioversion. They found a statistically significant, lower treatment failure rate for PFA than thermal ablation (OR 0.83, 95% CI 0.70 to 0.98, p=0.03). Subgroup analyses indicated this finding was not driven by whether it was measured before 1 year or after 1 year (p>0.97), whether PFA was compared with RFA or CBA (p=0.22), or whether the studies were done in people with PAF or PersAF (p=0.63).

In the ADVENT RCT that compared PFA with RFA and CBA (Farawave, Reddy 2023a, Patel 2024), composite score of treatment success was freedom from:

By this definition, estimated probability of treatment success was 73% for PFA and 71% for TA. Probability of treatment success for PFA was not statistically inferior to TA (posterior probability was more than 0.999). The difference did not reach the threshold to show superiority (posterior probability=0.708). All people had PAF in this study.

In MANIFEST-PF (Farawave), the EU retrospective analysis of a prospective registry of 1,568 people across 24 centres (Turagam 2023, Turagam 2024), Kaplan-Meier estimates of freedom from AF, atrial flutter or atrial tachycardia and freedom from antiarrhythmic drugs, or any need for redo ablation at 1 year was 71% (95% CI 68 to 73%). This outcome was statistically significantly better for people with PAF (74%) than people with PersAF (65%, p=0.001).

In PULSED-AF (PulseSelect), the international prospective single-arm trial of 300 people across 41 centres (Verma 2023a, Verma 2023b), treatment success was defined as freedom from acute procedural failure (inability to isolate all PVs, or any left atrium lesion using a non-study device during the index procedure), arrhythmia recurrence or antiarrhythmic escalation at 12 months, excluding a 90 day blanking period. A more detailed checklist of criteria is listed in the publication. In the PAF group, 67% of people met the primary effectiveness endpoint and the 1-year Kaplan-Meier estimate was 66% (95%CI 58 to 73%). In the PersAF group, 55% of people met the primary effectiveness endpoint and the 1-year Kaplan-Meier estimate was 55% (95% CI 47 to 63%). Recurrence of atrial fibrillation was the most common reason for classification of treatment failure in both groups. No statistical comparison was made between PAF and PersAF.

Recurrence of AF
Recurrence of AF was reported in 3 studies. In the systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024), AF recurrence was statistically significantly lower in people who had PFA compared with TA (RR 0.66, 95% CI 0.51 to 0.87, p=0.003). No other comparative evidence with TA was included in the key evidence.

In inspIRE (Varipulse), the EU and Canada prospective single-arm trial of 226 people across 13 centres (Duytschaever 2023, de Potter 2024, Grimaldi 2023), people in the wave 1 cohort had 12-month freedom from AF recurrence Kaplan-Meier estimates of 77% (95% CI 64 to 90%, Duytschaever 2023). In the most complete wave 2 (pivotal phase study cohort, de Potter 2024), 82% of 186 people had 12-month freedom from AF recurrence (95% CI 76.1 to 87.2%). Simulation modelling of real-world outcomes gave even higher estimates for freedom from all recurrence (86%) and freedom from symptomatic recurrence (94%). The increase in freedom from recurrence rates indicates there may be a learning curve effect as procedural workflows and operator experience becomes established.

In the PULSED-AF (PulseSelect) international prospective single-arm trial of 300 people across 41 centres (Verma 2023a, Verma 2023b), AF recurrence was reported separately for people with PAF and people with PersAF. No statistical comparison was made but recurrence of atrial fibrillation was the most common reason for classification of treatment failure in both groups, with similar proportions. AF recurrence rate was 20% in the PAF cohort and 26% in the PersAF cohort.

Recurrence of atrial arrhythmia
Recurrence of atrial arrhythmia, not just AF, was reported in 7 studies.

In the systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024) there was no statistically significant difference in atrial arrhythmia recurrence (RR 0.92, 95% CI 0.74 to 1.13, p=0.42). Other comparative evidence from the UK real-world study of 707 people comparing RFA with PFA and CBA (Farawave, Calvert, 2024) had similar findings. Kaplan-Meier estimates showed no statistically significant difference between PFA and TA in rate of 12-month arrhythmia freedom after a 2-month blanking period (p=0.34), and adjusted HRs of comparisons between PFA and CBA and RFA also found no statistically significant difference (p=0.177, p=0.417 respectively).

Across all studies, the highest rate of recurrence of atrial arrhythmia was in the UK real-world study of 707 people (Farawave, Calvert, 2024). In this study, 51% of people that had PFA had freedom from arrhythmia at 12 months. This was only after a 2-month blanking period whereas most studies defined recurrence after a 3-month blanking period. The lowest recurrence of arrhythmia outcomes were reported in the prospective case series of 191 people (Farawave, Schmidt 2022). This study reported that 6% of people had symptomatic recurrence of atrial tachyarrhythmia, although this was among only 69 people and reported at 6 months, whereas most studies reported this at 12 months. Also, this was symptomatic recurrence not any detected recurrence, as in some other studies.

Among the other 4 studies that reported it, atrial arrhythmia freedom rates ranged between 62% in people with PersAF (any symptomatic or asymptomatic atrial arrhythmia detected in the PULSED-AF trial- PulseSelect, Verma 2023a) to 85% in people with PersAF or PAF who had an optimised treatment protocol (Reddy 2023b, a publication that combined 2 prospective single-arm trials totalling 178 people in 3 EU centres). Findings across the 4 studies varied based on some factors. When PAF and PersAF were compared, people with PAF tended to have better outcomes (see comparisons in MANIFEST-PF- Farawave, Turagam 2023, Turagam 2024 and PULSED-AF- PulseSelect, Verma 2023a,b). Outcomes were more favourable if only symptomatic arrhythmia recurrence, rather than any arrhythmia detected with Holter monitoring was considered (see comparisons in PULSED-AF- PulseSelect, Verma 2023a, Verma 2023b).

Health-related quality of life
Two studies reported health-related quality of life outcomes. The ADVENT RCT (Farawave, Reddy 2023a, Patel 2024) found no difference in health-related quality of life outcomes between PFA and TA at 1 year. This was across 3 patient-reported measures, including 1 AF specific questionnaire (AFEQT). No statistical test was reported. In PULSED-AF, the international prospective single-arm trial of 300 people across 41 centres (PulseSelect, Verma 2023a), people with both PAF and PersAF showed statistically significant and clinically meaningful improvements in quality of life at 12 months compared to baseline. This was across 2 patient-reported measures, including 1 AF specific questionnaire (AFEQT). Additional analyses in Verma (2023b) found that most benefit in people with PAF was seen if they had no atrial arrhythmia burden at 12 months, and clinically relevant improvements were driven by people with PersAF and people with PAF with less than 10% atrial arrhythmia burden.

Reintervention
Six studies reported reintervention rates. Lowest rates were reported in the prospective case series of 191 people (Farawave, Schmidt 2022), in which 5 of 191 people had a repeat procedure. But, follow-up was only until 6 months in this study and most studies measured this at 12 months. Low rates of need for reintervention were also reported in the ADVENT RCT (Farawave, Reddy 2023a, Patel 2024), in which 5% of people who had PFA and 7% of people who had TA had reintervention for clinical recurrence of AF. Not all studies restricted reintervention outcomes just to reintervention for AF, some included reintervention for all atrial arrhythmia, which may explain higher rates of reintervention in other studies compared to ADVENT. Highest reintervention rates were reported in the UK real-world study of 707 people comparing RFA with PFA and CBA (Farawave, Calvert, 2024), in which 19% of people who had PFA had reintervention at a median of 9 months. This was similar to rates in the comparison arms of this study, with 18% of people who had CBA and 16% of people who had RFA also having reintervention.

The remaining 3 studies reported reintervention rates of between 8 and 9% up to 12 months (MANIFEST-PF- Farawave, Turagam 2023, Turagam 2024, inspIRE- Varipulse, Duytschaever 2023, de Potter 2024, Grimaldi 2023, and PULSED-AF- PulseSelect, Verma 2023a, Verma 2023b).

Acute procedural success
Acute procedural success was reported in 8 studies and defined in terms of acute isolation of target lesion sites. The systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024) found no statistically significant difference between PFA and thermal ablation (OR=1.62, p=0.64). Any signal of difference was likely driven by comparisons with RFA. There was no statistically significant difference in procedural success rates between PFA and CBA (p=0.61). All studies comparing PFA with RFA had 100% success rate in the PFA arm.

In contrast, the UK real-world study of 707 people comparing RFA with PFA and CBA (Farawave, Calvert, 2024) found acute success was statistically significantly higher in PFA (99%) and RFA (99%) than CBA (95%, p=0.008).

In the non-comparative studies, lowest rates of PVI without acute reconnection was seen in 96% of veins in wave 1 and 97% of veins in wave 2 of inspIRE (the EU and Canada prospective single-arm trial of 226 people across 13 centres- Duytschaever 2023, de Potter 2024, Grimaldi 2023). Rates were between 99% and 100% in all other studies.

Procedure time
Procedure time was reported in 10 studies. Both systematic review and meta-analyses (de Campos 2024, Amin 2024) found PFA procedure time was statistically significantly shorter than thermal ablation (MD -21.7 minutes, p<0.01 and MD -15.2 minutes, p<0.0001 in each study respectively), although heterogeneity was high. The de Campos (2024) subgroup analysis found that the effect size was probably driven by comparisons with RFA; the mean procedure time was 41 minutes longer in the RFA arm than PFA (p<0.01). Other comparative studies found PFA procedure time to be shorter than TA comparators (ADVENT RCT- Farawave, Reddy 2023a Patel 2024; UK real-world study of 707 people- Farawave, Calvert 2024). Average procedure time across all studies that reported it ranged from 39 minutes (SD 14) as reported in the prospective case series of 191 people (Farawave, Schmidt 2022), to 140 minutes (SD 42, CENTAURI, Anić, 2023).

Procedure time may be affected by technology characteristics ('point-by-point' PFA may take longer than 'single-shot' catheters), learning curve and procedure protocol establishment (see Schmidt 2022, inspIRE- Duytschaever 2023 and also Anić 2023), use of sedation or GA (see inspIRE- de Potter 2024, Grimaldi 2023) use of additional lesion sets (see MANIFEST-PF- Turagam 2023, Turagam 2024), or procedures done in people with PAF compared with PersAF (see PULSED-AF- Verma 2023a, Verma 2023b).

Length of stay
One study, the UK real-world study of 707 people comparing RFA with PFA and CBA (Farawave, Calvert, 2024), reported length of stay outcomes. They found that people who had CBA had a higher rate of same day discharge (59%) than people who had PFA (45%) or RFA (25%) (p<0.001). The authors explained this may be because more PFA and RFA procedures were done in the afternoon than CBA.

Use of additional lesion sets for linear ablation and outcomes
Many studies included people who had additional lesion sets at the operator's discretion using PFA in addition to PVI, but only 3 studies reported data relating to this.

In terms of prevalence of this practice in the UK, the UK real-world study of 707 people comparing RFA with PFA and CBA (Farawave, Calvert, 2024) reported that posterior wall isolation was done in 19 PFA procedures (9%) and 2 RFA procedures (2%).

A sub study of MANIFEST-PF (Farawave, EU retrospective analysis of a prospective registry of 1,568 people across 24 centres) did an analysis in people with PersAF or longstanding PersAF (Turagam 2024). Time from first to repeat ablation did not differ between the group who had left atrial posterior wall ablation in addition to PVI (220 days) and the PVI only group (202 days, p=0.92). Also, this study found that likelihood of repeat ablation did not differ between people who had left atrial posterior wall ablation (14%) and people who had PVI only (10%, p=0.26). They reported that there was no statistically significant difference in 1-year Kaplan-Meier estimates of freedom from atrial arrhythmias in people who had PVI plus left atrial posterior wall ablation (66%, 95% CI 58 to 74%) compared with people who had PVI only (73%, 95% CI 69 to 77%). A propensity score-matched cohort sensitivity analysis of 184 people also found no statistically significant difference on this outcome (p=0.34). There was also no statistically significant difference in median time to recurrence (PVI plus left atrial posterior wall 207 days compared with 178 days for PVI only, p=0.68).

In the publication that combined evidence from 2 prospective single-arm trials which together had 178 people across 3 EU centres (Sphere-9, Reddy 2023b), linear lesion block was reported. Because various combinations of RFA and PFA were used for most linear lesions, only the roof line data can be used to understand PFA effects alone. On roof line ablations, durability was between 63% for early waveforms to 100% with the optimised waveform.

Safety

Key safety events are presented below. Other minor events and events that were reported but not discussed in reference to the procedure are presented in table 3.

Overall complication rate
Both systematic review and meta-analyses reported overall complication rates. Both studies compared PFA with TA (de Campos 2024, Amin 2024), and in primary analyses, neither found any statistically significant difference in complication rates. In de Campos (2024), the OR was 0.79 (95% CI 0.47 to 1.33, p=0.38, I² =37%). But, sensitivity analyses did result in a statistically significant difference; heterogeneity reduced to 0% and fewer periprocedural complications were reported in the PFA than thermal ablation arms of the studies in 1 leave-one-out scenario (OR 0.59, 95% CI 0.39 to 0.88, p<0.01). In Amin (2024), no statistically significant difference was seen between PFA and TA (RR 0.9, 95% CI 0.80 to 1.02, p=0.10, 9 studies, I²= 17%).

Major complications (SAEs)
Four primary research studies done with the Farawave catheter and 1 with the Sphere-9 catheter reported summaries and comparisons of SAE rates.

Comparative studies found PFA to be either non-inferior or to have no statistically significant difference in SAE rate compared with TA (RFA and CBA) (Farawave- Reddy 2023a, Patel 2024, Calvert 2024). The rate of SAEs in the PFA arm ranged from 0.5% at a median of 7.6 months in the UK real-world study (Calvert, 2024), to 7% at 12 months for procedure-related events in the ADVENT RCT (Reddy 2023a, Patel 2024).

The ADVENT RCT (Reddy 2023a, Patel 2024) reported procedure-related and device-related SAEs for the Farawave catheter at 12 months separately. Device SAE rate was lower than procedure-related SAE rate for both PFA and TA (procedure related: 7% for PFA and 3% for TA; device related: 1% for PFA and no events for TA).

In the 2 large studies based mainly in the EU, similar event rates were reported with the Farawave catheter, up to 12 months. In the full analysis of MANIFEST-PF (Turagam 2023), 2% had a major complication. In MANIFEST-17K (Ekanem 2024) 1% had a major complication, which were most commonly vascular issues.

The publication that combined evidence from 2 prospective single-arm trials which together had 178 people across 3 EU centres (Sphere-9, Reddy 2023b) reported 4 procedure-related SAEs (2%).

Oesophageal lesions and injury
Eight studies reported incidence of oesophageal lesions, injury and atrio-oesophageal fistula. The primary studies were done with Farawave, PulseSelect, Sphere-9 and CENTAURI.

Both systematic reviews found lower oesophageal lesion rates in PFA than TA. In the systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024), there were statistically significantly lower rates of peri-oesophageal and oesophageal injuries in the PFA group than in the thermal ablation group (OR 0.17, 95% CI 0.06 to 0.46, p<0.01). In the systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024), incidence of oesophageal lesions was statistically significantly lower in PFA than TA (RR 0.09, 95 CI% 0.01 to 0.69, p=0.02).

There were no instances of oesophageal injury or complications in 3 studies done with Farawave and 1 done with PulseSelect. In MANIFEST-PF, the EU retrospective analysis of a prospective registry of 1,568 people across 24 centres (Farawave, Turagam 2023) reported there were no oesophageal complications, including no atrio-oesophageal fistula, oesophageal ulcerations or oesophageal dysmotility disorders. In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Farawave, Ekanem 2024), there were no incidents of oesophageal complications including no atrio-oesophageal fistula or dysmotility disorders. In the prospective case series of 191 people (Farawave, Schmidt 2022), 52 people had oesophageal endoscopy and there were no incidents of thermal injury. In PULSED-AF, the international prospective single-arm trial of 300 people across 41 centres (PulseSelect, Verma 2023a, Verma 2023b), there were no reports of atrio-oesophageal fistula, or other oesophageal injury. Optional oesophageal temperature monitoring demonstrated a 0.3 (plus or minus 0.4 degrees) Celsius change from baseline (140 people).

Mild oesophageal lesions were seen on invasive post-procedure monitoring in 2 studies. In the publication that combined evidence from 2 prospective single-arm trials done with the Sphere-9 catheter (178 people across 3 EU centres; Reddy 2023b) low level heating (around 2.3 degrees Celsius) was seen when PFA was done along the left atrial posterior wall. Post-procedure oesophagogastroduodenoscopy showed 3 instances (8% of 36 people) of asymptomatic minor mucosal thermal injury in a subgroup of people who had RFA and PFA. There were no injuries in the PFA only group. In the prospective, single-arm trial done with the CENTAURI generator in 84 people from 2 EU centres (Anić 2023) 3% (2 people of 75) had mild oesophageal lesions, seen on endoscopy within 72 hours, but these were the only people who also had oesophageal temperature probe placed during the procedure. No temperature rise was seen in relation to PFA and both events were resolved without sequalae.

Phrenic nerve injury
Seven studies reported on phrenic nerve injury. Most data was from the Farawave catheter and 1 study was done with the CENTAURI generator.

The systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024) found the incidence of phrenic nerve palsy was statistically significantly lower in PFA than TA (RR 0.38, 95% CI 0.15 to 0.98, p=0.05).

In 2 comparative studies with the Farawave catheter, no instances of phrenic nerve injury were seen in the PFA arm but these were seen in the CBA arm (ADVENT RCT Reddy 2023a, Patel 2024 and the UK real-world study - Calvert, 2024).

Rates between 0 and 1% were reported in the single-arm data collected in studies done with Farawave and CENTAURI. No phrenic nerve injury was seen in 80 people who were monitored in the prospective, single-arm trial done with the CENTAURI device in 84 people from 2 EU centres (Anić 2023). In MANIFEST-PF, the EU retrospective analysis of a prospective registry of 1,568 people across 24 centres (Farawave, Turagam 2023, Turagam 2024), 1 person had persistent phrenic nerve injury. This was classed as a major complication. In Turagam (2024), transient phrenic nerve injury happened in 1 person who had PVI plus left atrial posterior wall ablation. In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Farawave, Ekanem 2024), there were no reported instances of sustained phrenic nerve palsy but 0.1% (11 of 17,642 people) had transient phrenic paresis that recovered by the next day. In the prospective case series of 191 people (Schmidt 2022) 1% (2 of 191 people) had phrenic nerve palsy with loss of diaphragmatic contraction; 1 was in the validation phase, 1 was in the streamline phase of the study. In both cases, phrenic nerve function recovered within 1 minute.

Pericardial tamponade, perforation and effusion
Eleven studies reported on pericardial tamponade, perforation and effusion, across all technologies included in the prioritised evidence.

Both systematic review and meta-analyses found that rate of tamponade was statistically significantly higher than TA. The systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024) reported OR of 2.98 (95% CI 1.27 to 7.00, p=0.01; I²=0%). The systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024) reported a RR of 6.14 (95% CI 1.43 to 26.3, I²= 0%).

Among the studies with the Farawave technology, tamponade rates ranged from 0 (in the UK real-world study - Calvert, 2024) to 1% (in MANIFEST-PF, Turagam 2023). Some events did results in complications and death. The ADVENT RCT (Reddy 2023a, Patel 2024) reported pericardial tamponade was seen in 2 of 305 people who had PFA (0.6%). In 1 of the 2 people, this was fatal. In MANIFEST-PF (Turagam 2023, Turagam 2024), tamponade rate was 1%. Turagam (2023) reported this was the most common major AE. In both studies some people required surgical drainage to treat symptoms. Pericardial effusion, classed as a minor AE and without need for intervention, was also seen in 0.3% of people in this study. In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Ekanem 2024), 0.4% (63 of 17,642 people) had pericardial tamponade. Most instances (56 of 63 people) were managed with percutaneous pericardiocentesis. The other 7 people had surgery and 1 person died from complications. In the prospective case series of 191 people (Schmidt 2022), 1 person had cardiac perforation with pericardial effusion, in the streamline phase of the study. They had pericardiocentesis and anticoagulation reversal and were discharged within 2 days.

In inspIRE (Varipulse), the EU and Canada prospective single-arm trial of 226 people across 13 centres (Duytschaever 2023, de Potter 2024, Grimaldi 2023), there were no incidents of cardiac tamponade or perforation.

In PULSED-AF (PulseSelect), the international prospective single-arm trial of 300 people across 41 centres (Verma 2023a, Verma 2023b) 1 person with PersAF had pericardial effusion needing drainage.

In the publication that combined evidence from 2 prospective single-arm trials which together had 178 people across 3 EU centres (Sphere-9, Reddy 2023b) there was 1 pericardial effusion event related to transeptal puncture. But, the authors reported no events of tamponade or perforation.

In the prospective, single-arm trial of 84 people from 2 EU centres (CENTAURI, Anić 2023) 2 events of catheter perforation were seen in 2 people, which led to cardiac effusion or tamponade. In 1 person, it was resolved with pericardiocentesis and they were discharged after an overnight stay. In 1 person, the effusion and tamponade was resolved with cardiac surgery. The event was resolved but because of the delay in treatment it caused a non-embolic cerebrovascular event (this was also recorded as a primary SAE endpoint).

Stroke, transient ischaemic attack
Eight studies reported on stroke and transient ischaemic attack events, across the Farawave, PulseSelect and Sphere-9 technologies.

In the systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024), there was no statistically significant difference between PFA and TA in the incidence of stroke or transient ischaemic attack (RR 0.52, 95% CI 0.14 to 1.91, p=0.32).

Absolute estimates of the rate and transient ischaemic attack were low, between 0 and 1% but did sometimes result in death. No events happened in the PFA groups reported in the ADVENT RCT (Reddy 2023a, Patel 2024), UK real-world study (Calvert, 2024) or the publication that combined evidence from 2 prospective single-arm trials with 178 people across 3 EU centres (Reddy 2023b). In MANIFEST-PF, the EU retrospective analysis of a prospective registry of 1,568 people across 24 centres (Turagam 2023, Turagam 2024), stroke was the second most common major complication. This happened in 0.4% (11 of 1,568 people). They also reported that 1 person died during the study from a sustained stroke. In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Ekanem 2024), 0.1% (22 of 17,642 people) had stroke. The most common cause was catheter exchanges or sheath management in the events that a cause could be determined. In the prospective case series of 191 people (Schmidt 2022), 2 people had stroke. They were both in the validation phase of the study. In 1 person, symptoms had resolved within 4 weeks. In the other person, symptoms had resolved by 3 months. In PULSED-AF, the international prospective single-arm trial of 300 people across 41 centres (Verma 2023a, Verma 2023b), 1 person (of 150) with PAF had a cerebrovascular accident on the same day as the procedure. They had lower leg numbness and mild dysphagia. It was resolving at the end of the study.

Silent cerebral events
Seven studies reported findings on silent cerebral events and lesions. These tended to be done in subgroups of the full study sample and studies were done with the Farawave, Varipulse, PulseSelect, Sphere-9 and CENTAURI technologies.

Across all 6 studies, silent cerebral event rates were very high in some analyses of some studies. Comparative evidence from the ADVENT RCT (Reddy 2023a, Patel 2024) found no statistically significant difference in rate of silent cerebral events and lesions detected on MRI when compared to TA (p=0.10). Neurological findings for all 3 people who had silent cerebral events in the PFA arm were normal at predischarge and day 90 assessments, and mRS and NIHSS scores were 0 at each timepoint indicating no neurological deficit. In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Farawave, Ekanem 2024), a subset of 96 people at 8 centres had routine post-procedural MRI. Abnormalities were seen in 9 people (9%). In the prospective case series (Varipulse, Schmidt 2022) 53 of 191 people in the study had diffusion weighted MRI within 24 hours of the ablation. Acute asymptomatic injury was seen in 19% (10 people). All of these procedures were done in the procedure development phase of the study. None of the people had symptoms. In inspIRE (Farawave), the EU and Canada prospective single-arm trial of 226 people across 13 centres (Duytschaever 2023, de Potter 2024, Grimaldi 2023), silent cerebral lesions were found 4 of the first 6 of the wave 1 cohort. After workflow enhancements, 4 silent cerebral lesions were found in the following 33 people. All lesions were asymptomatic and transient (resolved by 3 months). All participants had National Institutes of Health Stroke Scale score of 0 by 3-month follow-up. In PULSED-AF (PulseSelect), the international prospective single-arm trial of 300 people across 41 centres (Verma 2023a, Verma 2023b), 45 people had cerebral MRI before and after the procedure and 4 of 45 people (9%) showed a new silent cerebral lesion. But, change in Mental State Examination scores only changed by 0.4 (SD 1.8). In the publication that combined evidence from 2 prospective single-arm trials centres (Sphere-9, Reddy 2023b), 8% (7 of 89 people) had diffusion weighted imaging-positive but fluid attenuated inversion recovery-negative silent cerebral events and 7% (6 of 89 people) had diffusion weighted imaging-positive and fluid attenuated inversion recovery-positive silent cerebral lesions at a mean of 1.2 days post-procedure. In the prospective, single-arm trial of 84 people from 2 EU centres (CENTAURI, Anić 2023), a subgroup of 35 people had MRI within 72 hours post-procedure. People with abnormal findings at 72 hours returned for 30 day follow up MRI. 11% of people had silent cerebral events detected on diffusion weighted imaging (4 of 35 people), but not on fluid-attenuated inversion recovery (0%). All findings were resolved at 30 days with no incidence of neurological dysfunction, increase in National Institute of Health Stroke Scale or detectable microbubbles.

Systemic embolic events
Both systematic review and meta-analyses and 1 primary study reported on systemic embolic events. The systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024) found no statistically significant difference between PFA and thermal ablation in systemic embolic events (12 studies, OR 1.52, 95% CI 0.57 to 4.07, p=0.40). Similarly, the systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024) found there was no statistically significant difference between PFA and TA in the incidence of systemic thromboembolism (RR 0.33, 95% CI 0.01 to 8.01, p=0.50, 6 studies).

The publication that combined evidence from 2 prospective single-arm trials done in the Sphere-9 catheter, which together had 178 people across 3 EU centres (Reddy 2023b) reported that no thromboembolic events were seen.

Myocardial infarction
In the UK real-world study of 707 people comparing RFA with PFA and CBA (Farawave, Calvert, 2024) there was 1 suspected myocardial infarction event in the CBA arm. No events were seen in the PFA or RFA arm.

Three studies done with Farawave, CENTAURI and Sphere-9 reported no myocardial infarction events were seen (MANIFEST-17K- Ekanem 2024, Anić 2023, Reddy 2023b).

PV stenosis
Six studies reported on PV stenosis across the Farawave, PulseSelect, Sphere-9 and CENTAURI technologies. All 5 studies reported no evidence of PV stenosis. Some evaluated this in terms of symptomatic PV stenosis whilst others did invasive mapping.

The ADVENT RCT comparing PFA with RFA and CBA (Reddy 2023a, Patel 2024) reported that no people in either the PFA or TA groups had symptoms of pulmonary vein stenosis. Mean change in pulmonary vein cross-sectional area was superior in PFA to TA (mean relative difference 11%, 95% CI 8 to 14%, posterior probability of superiority >0.999). In MANIFEST-PF, the EU retrospective analysis of a prospective registry of 1,568 people across 24 centres (Turagam 2023, Turagam 2024), they reported there were no incidents of PV stenosis. In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Ekanem 2024), there were no reported events of PV stenosis. In PULSED-AF, the international prospective single-arm trial of 300 people across 41 centres (Verma 2023a, Verma 2023b) there were no moderate or severe pulmonary vein stenosis was observed on imaging between baseline and 3 months (63 people). In the publication that combined evidence from 2 prospective single-arm trials across 3 EU centres (Reddy 2023b), there was no evidence of any PV stenosis in all 122 people who had invasive remapping. Cardiac computed tomography scans were done on everyone at 129 days (SD 93); 43% (77 of 178 people) indicated no evidence of PV stenosis. In the prospective, single-arm trial of 84 people from 2 EU centres (Anić 2023), they found that between baseline and 90 days, mean narrowing was between 0.5% and 3.6% across PVs which indicated no chronic effect.

Coronary artery spasm
Both MANIFEST studies done with Farawave reported coronary spasm events. In MANIFEST-PF, (Turagam 2023, Turagam 2024), coronary spasm happened in 0.1% (2 of 1,568 people). This was classed as a major complication. This included 1 person who had PVI plus left atrial posterior wall ablation. Both events were resolved with intracoronary nitroglycerin. In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Ekanem 2024), 0.1% (25 of 17,642 people) had coronary spasm. Most were proximal to the lesion (88%), 20% (5 of 25 people) had hypotension and nitroglycerin was given to 84% (21 of 25 people). Clinical sequalae were seen in 4 people; 1 person had atrioventricular block and ventricular fibrillation during PFA of the cavotricuspid isthmus and they had resuscitation and defibrillation, 2 people had chest pain in the post-procedure recovery area and both resolved with nitroglycerin, and 2 people had anterior ST elevation, polymorphic premature ventricular contractions and subsequent ventricular fibrillation after PFA at the right inferior PV. They had resuscitation, defibrillation and intravenous nitroglycerin.

In PULSED-AF (PulseSelect), the international prospective single-arm trial of 300 people across 41 centres (Verma 2023a, Verma 2023b), they reported there were no coronary artery spasm events.

Pericarditis
Four studies reported pericarditis outcomes across the Farawave, Varipulse and CENTAURI technologies.

In the systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024), they reported that across the 18 included studies (with 1,761 people who had PFA), 2 events of pericarditis were reported.

In MANIFEST-PF (Turagam 2023, Turagam 2024), 1 person had pericarditis classed as a minor complication.

Both inspIRE (Varipulse), the EU and Canada prospective single-arm trial of 226 people across 13 centres (Duytschaever 2023, de Potter 2024, Grimaldi 2023), and the prospective, single-arm trial of 84 people from 2 EU centres (CENTAURI, Anić 2023) reported there were no pericarditis events.

Haemoglobinuria, haemolysis and acute kidney injury
Two studies reported on acute kidney injury and associated safety events. They were both done with Farawave.

One study explicitly reported differences in haemoglobinuria, haemolysis and acute kidney injury depending on periprocedural hydration (Mohanty, 2024). In this study, group 1 (28 people) had little or no hydration immediately after the procedure and group 2 (75 people) had planned fluid infusion. Procedures in this study were done with the Farawave technology.

In group 1, 75% had haemoglobinuria within 24 hours post-PFA and there was a statistically significant increase in bilirubin, indicating haemolysis (p=0.002). Also, all 21 people with haemoglobinuria had statistically significantly increased serum creatinine levels post-ablation (p<0.001), and 19 people had serum creatinine levels that met diagnostic criteria for acute kidney injury. All 4 with renal kidney injury had haemoglobinuria after the procedure and had statistically significantly more PFA applications than the other 17 people (p<0.001). Serum creatinine was in the normal range for people without haemoglobinuria, and they had statistically significantly fewer PFA applications than people who had haemoglobinuria (p=0.019). All people with haemoglobinuria were successfully treated with hydration before discharge.

In group 2, there was no incidence of haemoglobinuria, no statistically significant change in bilirubin (p=0.62) and no statistically significant change in serum creatinine (p=0.18).

In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Ekanem 2024), 5 people (0.3%) had haemolysis with acute kidney failure with creatine level increase of 100% within a day of the procedure. Symptoms included haemoglobinuria, nausea and oliguria, beginning either immediately post-procedure or the next day. Transient haemodialysis improved renal function by the time of hospital discharge. All 5 had PFA for PersAF, with a complex lesion set with a high number of PF applications. One other person had haemolysis without acute kidney injury and the authors reported that several people at 1 centre had 'dark urine' or haemoglobinuria either immediately post-procedure or the next day. But there was no reported kidney injury or drop in red cell count in these people.

High sensitivity troponin level
The systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024) reported a summary of high-sensitivity troponin levels, which can indicate heart damage. They found results were statistically significantly higher in PFA than TA in the sensitivity analysis (3 studies, MD 421.4, 95% CI 251.5 to 591.4, p<0.01, I²=77%).

Vascular access complications
Eight studies reported findings on vascular access complications across the Farawave, Varipulse, Sphere-9 and CENTAURI technologies.

The systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024) found no statistically significant difference between PFA and thermal ablation in vascular access complications (OR 0.91, 95% CI 0.54 to 1.54, p=0.73).

Among studies done with Farawave, vascular access complications happened in 2% or less of people. In the UK real-world study of 707 people comparing RFA with PFA and CBA (Calvert, 2024), there were 3 serious vascular injuries but only 1 in the RFA arm, which needed surgical intervention. In MANIFEST-PF, the EU retrospective analysis of a prospective registry (Turagam 2023, Turagam 2024), 0.1% (2 of 1,568 people) had vascular access complications requiring surgery. This was classed as a major complication. In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Ekanem 2024), 3% (441 of 17,642 people) had vascular complications, of which 0.3% were major complications. Rate of major vascular complication was statistically significantly higher in centres that did not routinely use ultrasound guidance (p=0.046). In the prospective case series of 191 people (Schmidt 2022), 2% (4 of 191 people) had a vascular access complication; 2 were in the validation phase, 2 were in the streamline phase. These were major groin haematomas and resolved with conservative management.

In inspIRE (Varipulse), the EU and Canada prospective single-arm trial of 226 people across 13 centres (Duytschaever 2023, de Potter 2024, Grimaldi 2023), there were no primary adverse events which included major vascular access complication or bleeding.

In the publication that combined evidence from 2 prospective single-arm trials which together had 178 people across 3 EU centres (Sphere-9, Reddy 2023b), there were 2 vascular access SAEs, including 1 groin haematoma requiring surgical intervention and 1 groin puncture bleeding treated with compression. This study also reported there were no major vascular complications.

In the prospective, single-arm trial of 84 people from 2 EU centres (CENTAURI, Anić 2023), 3 vascular access complications led to haemorrhagic events needing hospitalisation.

Heart rate change and ST segment changes
The systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024) reported change in heart rate findings. PFA had significantly lower heart rate change than TA (MD -7.39 beats per minute, 95% CI -12.2 to -2.62, p=0.002, 4 studies, I²= 86%). Leave one out analysis resolved heterogeneity with removal of 1 study, after which MD was -9.78 (95% CI -11.8 to -7.7, p<0.00001, 3 studies I²=0%).

ST-segment change events were rare. The systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024) reported that across the 18 included studies (with 1,761 people who had PFA), there was 1 event of ST segment change. The publication that combined evidence from 2 prospective single-arm trials which together had 178 people across 3 EU centres (Reddy 2023b) reported 1 transient ST-segment elevation following atropine administration in a person with prior myocardial infarction and unknown residual right coronary arterial stenosis later requiring angioplasty and stenting. In MANIFEST-17K (Ekanem 2024), 2 people had anterior ST elevation, polymorphic premature ventricular contractions and subsequent ventricular fibrillation after PFA at the right inferior PV. This was considered a clinical sequalae to coronary spasm. They had resuscitation, defibrillation and intravenous nitroglycerin. Both PULSED-AF, the international prospective single-arm trial of 300 people across 41 centres (Verma 2023a, Verma 2023b) and the prospective, single-arm trial of 84 people from 2 EU centres (Anić 2023) found there were no incidences of ST elevation during any PFA applications.

Death
Ten studies reported mortality outcomes across Farawave, Varipulse, PulseSelect, Sphere-9 and CENTAURI technologies.

The systematic review and meta-analysis of 17 studies comparing PFA with TA (Amin 2024) found there was no statistically significant difference in all-cause mortality between PFA and TA (RR 0.33, 95% CI 0.01 to 8.07, 7 studies). The systematic review and meta-analysis of 18 studies comparing PFA with TA (de Campos 2024) did not do comparative analysis but reported 1 death among the 1,761 people who had PFA.

In 4 studies, the UK real-world study (Farawave, Calvert, 2024), the inspIRE EU and Canada prospective single-arm trial (Duytschaever 2023, de Potter 2024, Grimaldi 2023), the publication that combined evidence from 2 prospective single-arm trials (Sphere-9, Reddy 2023b) and the prospective single-arm trial (CENTAURI, Anić 2023) no deaths were reported.

In the ADVENT RCT (Reddy 2023a, Patel 2024) 1 person in the PFA arm had pericardial tamponade; emergency sternotomy and resuscitation were needed and the person died after 10 days. In MANIFEST-PF, the EU retrospective analysis of a prospective registry (Farawave, Turagam 2023, Turagam 2024), 1 person died during the study. This person had a sustained stroke. In MANIFEST-17K, the centre-level study of 101 centres with 17,642 people (Ekanem 2024), 0.03% (5 of 17,642 people) died during the study of which 2 deaths were classed and procedure-related: 1 person died from complications from cardiac tamponade, and 1 person died from post-procedure cardiogenic shock in a person with cardiomyopathy and decompensated heart failure. In PULSED-AF (PulseSelect), the international prospective single-arm trial of 300 people (Verma 2023a, Verma 2023b) 1 person with PAF died during follow-up who had a history of cirrhosis, a CT scan showed liver masses and they died of liver failure and 1 person with PersAF died from cardiopulmonary arrest within 2 weeks of dofetilide initiation.

Composite safety outcomes
Four studies defined a primary safety endpoint which could be met if one of a selection of predefined SAEs was seen. These were in studies done with 4 technologies (Varipulse, PulseSelect, Sphere-9 and CENTAURI). Authors reported that events were selected as part of the composite because they are known procedure or device-related events, although their definitions were not homogeneous across studies. Between 0 and 4 people met safety endpoints in each of the 4 studies.

In inspIRE (Varipulse), the EU and Canada prospective single-arm trial of 226 people across 13 centres (Duytschaever 2023, de Potter 2024, Grimaldi 2023), there were no primary adverse events (defined as pericarditis, myocardial infarction, cardiac tamponade or perforation, thromboembolism, stroke or cerebrovascular accident, transient ischemic attack, phrenic nerve paralysis, or major vascular access complication or bleeding, as well as death, PV stenosis, and atrio-oesophageal fistula that occurred later than 7 days post-procedure).

In PULSED-AF (PulseSelect), the international prospective single-arm trial of 300 people across 41 centres (Verma 2023a, Verma 2023b), the primary safety endpoint was a composite of serious procedure and device-related AEs. One person with PAF and 1 person with PersAF of 150 people in each group met the primary safety endpoint at 6 months, and the 6-month Kaplan-Meier estimates were 0.7% (95% CI 0.1 to 4.6%) in both groups.

In the publication that combined evidence from 2 prospective single-arm trials of the Sphere-9 catheter, which together had 178 people across 3 EU centres (Reddy 2023b), 1 person met the primary safety endpoint. This was device-related serious adverse events within 7 days of the procedure. The person had combined RFA and PFA was hospitalised 20 days after the procedure for inflammatory pericardial effusion. It was managed with anti-inflammatory medication. There were no instances of death, myocardial infarction, persistent diaphragmatic paralysis, stroke or transient ischaemic attack or thromboembolism, cardiac tamponade or perforation, pneumothorax, major vascular complications, pulmonary oedema, hospitalisation or heart block. Note that PV stenosis, atrio-oesophageal fistula were not included in this assessment period.

In the prospective, single-arm trial of the CENTAURI generator, with 84 people from 2 EU centres (Anić 2023), the event rate was 5% (4 events in 4 of 82 people). The SAE endpoint was predefined system-related and procedure-related SAEs within 30 days of ablation. These included 3 vascular access complications led to haemorrhagic events needing hospitalisation and 1 non-embolic cerebrovascular accident caused by exacerbated cardiac tamponade secondary to ablation catheter perforation. There were no incidences of AE fistula, diaphragmatic paralysis, myocardial infarction, pericarditis, thromboembolism, PV stenosis, TIA, or death.

Differences in safety for PVI compared with PVI plus left atrial posterior wall ablation
In the substudy of MANIFEST-PF, (Turagam 2024), they found no statistically significant difference (p=0.51) in the major adverse event rate between people who had just PVI (2%; 3 of 131 people) compared with people who had PVI plus left atrial posterior wall ablation (1%; 6 of 416 people). The minor complication rate was 7% in people who had PVI plus left atrial posterior wall ablation and 4% in people who had PVI only. Transient phrenic nerve injury happened in 1 person who had PVI plus left atrial posterior wall ablation. Coronary spasm happened in 1 person who had PVI plus left atrial posterior wall ablation. Stroke happened in 1 person who had PVI plus left atrial posterior wall ablation and 1 person who had PVI only. No deaths happened in either group. As in the full trial analysis sample, there were no instances of PFA-related oesophageal complications, including no atrio-oesophageal fistula, oesophageal ulcerations, or oesophageal dysmotility. There were no instances of symptomatic PV stenosis or persistent phrenic nerve injury.

MHRA safety alert
An MHRA safety alert was issued for a series of lot numbers for Varipulse (Biosense Webster Inc) in April 2024. A polymer used to bind the catheter electrodes to the catheter (Polyurethane) had overrun over the electrodes.

In summary, the alert states that this may increase the risk of a cerebrovascular event. Five complaints were reported of people treated with Varipulse catheters from the listed lots. The report states that in all cases, the ablation procedures were completed successfully, and the patients were discharged home and the complaints are being investigated to determine if they are related to this issue. Healthcare providers who have treated people using the specific lots subject to this removal should continue to follow those patients pursuant to their standard of care.

Validity and generalisability

Any ongoing trials

UK registries
Professional experts informed NICE that there are ongoing UK registries (attached to the MINTAC Programme) dedicated to each technology in current NHS usage. The registries collect information on acute efficacy, safety, and resource utilisation, including a single 3-month follow up to look for late complications. At the moment, the Farawave technology is the most used in the NHS, with over 1,000 procedures done and collected in the MINTAC Farawave registry. Other technologies currently in this programme are PulseSelect and Sphere-9, Varipulse and CENTAURI. Experts on the CRM device working group said that compliance with the registries is a mandatory requirement for centres selected to have access to the technologies, and compliance is 100% to date. The registries are individual to each technology and do not have ethics committee oversight so cannot be recommended for use in NICE guidance. But, the group intend to publish reports on the data they are collecting.

NICOR were also contacted for data relating to PFA for AF. They do not currently have a field to indicate whether the ablative technique was done with PFA. So, this registry is currently inappropriate for NICE to recommend its use. But, NICOR have been made aware and expressed intentions to update the database.

Farapulse

NCT05072964 The FARAPULSE FARA-Freedom Trial A Prospective Open Label Single Arm Post Market Clinical Follow-Up Trial of the FARAPULSE Pulsed Field Ablation System in Patients With Paroxysmal Atrial Fibrillation https://clinicaltrials.gov/study/NCT05072964, n=180, EU. Expected completion: 07/2024

NCT05493852 A RWS of the FARAPULSE in A Chinese Population With PAF https://clinicaltrials.gov/study/NCT05493852 n=30, China. Expected completion: 31/12/2024

NCT06175234 Feasibility Study on the FARAVIEW Technology (NAVIGATE PF) https://clinicaltrials.gov/study/NCT06175234 Estimated n=30 Expected completion: 15/07/2024

NCT05159492 Ground-Breaking Electroporation-based Intervention for PAROXysmal Atrial Fibrillation Treatment (BEAT PAROX-AF) https://clinicaltrials.gov/study/NCT05159492 n=292, EU. Expected completion: 01/02/2025

NCT05443594 A Prospective Single Arm Open Label Study of the FARAPULSE Pulsed Field Ablation System in Subjects with Persistent Atrial Fibrillation (ADVANTAGE AF) https://clinicaltrials.gov/study/NCT05443594 n=669, US, Canada, EU. Expected completion: 03/2025.

NCT05418725 Ground-Breaking Electroporation-based Intervention for PERSistent Atrial Fibrillation Treatment (BEAT PERS-AF) https://clinicaltrials.gov/study/NCT05418725?term=NCT05418725&rank=1 , n=78, EU. Expected completion: 05/2025.

NCT06128174 LS-PersAFone: Pulsed Field Ablation for Long-Standing Persistent Atrial Fibrillation (LS-PersAFone) https://clinicaltrials.gov/study/NCT06128174?term=NCT06128174&rank=1

n=25, US. Expected completion: 06/2025

NCT05940597 Pulsed Field Ablation vs Cryoablation In Paroxysmal Atrial Fibrillation (FACIL AF) https://clinicaltrials.gov/study/NCT05940597 Estimated n=350, France. Expected completion: 09/2026

NCT06335082 A Registry Based Collaborative to Measure Efficiency, Effectiveness, and Safety of Farapulse PFA Technology for AF (DISRUPT-AF) https://clinicaltrials.gov/study/NCT06335082 n=2000, US. Expected completion: 30/01/2027

NCT05534581 Single Shot Pulmonary Vein Isolation: Comparison of Cryoballoon vs. Pulsed Field Ablation in Patients With Symptomatic Paroxysmal Atrial Fibrillation - A Multi-Center Non-Inferiority Design Clinical Trial (The SINGLE SHOT CHAMPION Trial) https://clinicaltrials.gov/study/NCT05534581 Estimated n=210, Switzerland. Expected completion: 01/2027

NCT05501873 Real World Data Collection in Subjects Treated With the FARAPULSE Pulsed Field Ablation System (FARADISE) https://clinicaltrials.gov/study/NCT05501873. n=1173, global including UK. Expected completion: 08/2027

NCT06096337 Pulsed Field Ablation (PFA) Vs Anti-Arrhythmic Drug (AAD) Therapy As a First Line Treatment for Persistent Atrial Fibrillation (AVANTGUARD) https://clinicaltrials.gov/study/NCT06096337 n=520, global. Expected completion: 12/2027

NCT06199180 Pulsed Field Ablation Versus Conventional Radiofrequency Catheter Ablation for Repeat PVI in Patients With Paroxysmal AF (REPEAT-AF) https://clinicaltrials.gov/study/NCT06199180 Estimated n=154, Netherlands. Expected completion: 07/2028

NCT06364215 Posterior Wall Substrate Modification Using Irreversible Electroporation for Paroxysmal Atrial Fibrillation (SMILE-AF) https://clinicaltrials.gov/study/NCT06364215, Estimated n=450, US. Expected completion: 10/2028

Varipulse

NCT06099730 Safety and Effectiveness of Pulmonary Vein Isolation and Posterior Wall Ablation with Pulsed Field Energy in Patients With Paroxysmal and Persistent AF (POLARIS) https://clinicaltrials.gov/study/NCT06099730 Estimated n=360, US. Expected completion: 11/2024

NCT06056557 Versatility of a Circular Multielectrode Catheter in the Individualized Recognition & Treatment of Atrial Fibrillation and Related Arrhythmias Using Pulsed Field Energy (VIRTUE) https://clinicaltrials.gov/study/NCT06056557 Estimated n=141, US. Expected completion: 31/12/2024

NCT05552963 A Study of Multi-electrode Circular Irreversible Electroporation (IRE) Catheter and Multi-Channel IRE Generator in Paroxysmal Atrial Fibrillation (AFIRE) https://clinicaltrials.gov/study/NCT05552963 Estimated n=147, China. Expected completion: 31/12/2024

NCT05971693 A Study For Treatment of Paroxysmal Atrial Fibrillation (PAF) With the OMNYPULSE Catheter and the TRUPULSE Generator (Omny-IRE) https://clinicaltrials.gov/study/NCT05971693?term=NCT05971693&rank=1 n=188, Canada, EU. Estimated completion: 04/2025

NCT06455098 A Study of Assessment on Safety and Effectiveness of BWI Pulsed Field Ablation With OMNYPULSE Catheter for the Treatment of Paroxysmal Atrial Fibrillation (PAF) (OMNY-AF) https://clinicaltrials.gov/study/NCT06455098 n=440, US, Australia. Estimated completion: 05/2026

NCT06014996 Comparison of PFA vs. RFA in Patients With Symptomatic Paroxysmal Atrial Fibrillation (InsightPFA) https://clinicaltrials.gov/study/NCT06014996 Estimated:292, China. Expected completion: 31/12/2026

Sphere-9

NCT05120193 Treatment of Persistent Atrial Fibrillation with Sphere-9 Catheter and Affera Mapping and Ablation System (SPHERE Per-AF). https://clinicaltrials.gov/study/NCT05120193 n=477, US and EU. Estimated completion: 10/01/2024.

Globe, Kardium

NCT05462145 Safety and Effectiveness of the Globe® Pulsed Field System for Treating Patients With Symptomatic Paroxysmal or Persistent Atrial Fibrillation (PULSAR) https://clinicaltrials.gov/study/NCT05462145 n=449, US, Canada, EU. Expected completion: 02/2025

THERMOCOOL STSF with TRUPULSE Generator

NCT06144632 A Study of the THERMOCOOL SMARTTOUCH Surround Flow (SF) Catheter With the TRUPULSE Generator for Treatment of Drug Refractory Symptomatic PAF (SMARTPULSE PAF). https://clinicaltrials.gov/study/NCT06144632 Estimated n=250, US. Expected completion: 16/06/2025

NCT06272981 A Study of the THERMOCOOL SMARTTOUCH Surround Flow (SF) Catheter With the TRUPULSE Generator for Treatment of Drug Refractory Symptomatic PAFOUS (PulseSmart) https://clinicaltrials.gov/study/NCT06272981 Estimated n=135, Australia, Canada. Expected completion: 29/08/2025

No technology named

NCT05328882 Early Recurrences of Atrial Arrhythmias and Their Impact on Late Recurrence After Pulmonary Vein Isolation With Pulsed Field Ablation of Paroxysmal Atrial Fibrillation https://clinicaltrials.gov/study/NCT05328882 Estimated n=40, Croatia. Expected completion: 17/05/2024

NCT06039722 Clinical Trial of the Pulse Field Ablation System for the Treatment of Paroxysmal Atrial Fibrillation in 11 Centers in China https://clinicaltrials.gov/study/NCT06039722 n=166, China. Expected completion: 17/08/2024

NCT06160076 Inflammatory Response Following " Pulsed Field Ablation " vs. Radiofrequency Ablation-2 (RIPAF-2) https://clinicaltrials.gov/study/NCT06160076 Estimated n=60, France. Expected completion: 16/12/2024

NCT06307860 Comparison of Different Ablation Surgeries on Left Atrial Reverse Remodelling in Patients With Atrial Fibrillation https://clinicaltrials.gov/study/NCT06307860

Estimated n=120, China. Expected completion: 01/03/2025

NCT05618340 PFA for Paroxysmal Atrial Fibrillation https://clinicaltrials.gov/study/NCT05618340, n=149, China. Expected completion: 01/09/2025

NCT06166524 Pulsed-field Ablation for Patients With Asymptomatic Non-paroxysmal Atrial Fibrillation https://clinicaltrials.gov/study/NCT06166524. Estimated n=124, Czechia. Expected completion: 31/12/2025