FormalPara Key Summary Points

Pulsed field ablation (PFA) is a novel, minimally thermal technology for treating paroxysmal atrial fibrillation (PAF), with effectiveness comparable to cryoballoon ablation (CBA) and low safety risk.

With limited direct evidence, the current study aimed to compare the safety profiles of PFA and CBA through indirect treatment comparison.

PFA was associated with significantly lower risk of major adverse events and prespecified adverse events (risk ratios of 0.16 and 0.53, respectively).

The findings of this study show that PFA has a more favorable safety profile than CBA, providing further insight into treatment strategies for patients with drug-refractory PAF.

Introduction

Atrial fibrillation (AF) is the most common type of cardiac arrythmia, affecting ~ 3% of the adult population across the world [1]. Atrial fibrillation is associated with a considerable impact on patient quality of life and substantial risk of cardiovascular events, leading to an increased risk of morbidity and mortality among patients [1, 2]. In addition, AF places considerable burden on the healthcare system [2].

Guidelines have recommended using anti-arrhythmic drugs (AADs) as an initial treatment of paroxysmal AF (PAF). Among patients who are drug-refractory or intolerant to AADs, pulmonary vein isolation (PVI) using catheter ablation is recommended. The use of catheter ablation has been increasing in its frequency over the past decade, with higher efficiency rates and better safety profile [3]. Cryoballoon ablation (CBA) is a thermal-based catheter ablation technology that uses cryo, or freezing, energy to freeze tissue and isolate the pulmonary veins to prevent aberrant electrical signals that are responsible for AF [4]. This technology has demonstrated lower recurrence rate and improved quality of life compared to AADs as a treatment of PAF [5]. Despite these clinical advantages, CBA is associated with increased risk of atrio-esophageal injury/fistula [6, 7], phrenic nerve injuries [8, 9], and/or pulmonary vein stenosis [10].

Pulsed field ablation (PFA) is a novel, minimally thermal catheter ablation technology that creates lesions in myocardial tissue through irreversible electroporation [11]. It is currently limited to high-volume centers, but its adoption rate continues to grow for smaller centers. Several single-arm clinical trials have shown the effectiveness of PFA in patients with drug-refractory PAF, with freedom from atrial arrhythmias ranging from 66.2% to 87.4% one year after PVI [12,13,14,15]. Because of the highly selective nature of PFA for myocardial tissue, PFA may prevent potential damage to adjacent tissues [16]. Evidence from clinical trials support favorable safety profiles of PFA, with incidence rates ranging from 0% to 2.5% for early-onset serious adverse events, including phrenic nerve injury, atrio-esophageal fistula, pulmonary vein stenosis, and vascular complications [12, 13].

Most clinical trials assessing PFA technology in patients with drug-refractory PAF are single-arm trials, and head-to-head trials comparing PFA to CBA are limited [17]. Indirect treatment comparisons pooling data from multiple clinical trials present a robust alternative option to compare these technologies. The objective was to conduct a critical literature review and an indirect treatment comparison to evaluate the comparative safety of PFA and CBA for PVI in patients with drug-refractory PAF.

Methods

Search Strategy

Literature searches were conducted in the MEDLINE (PubMed) database and the Clinicaltrials.gov registry from January 2008 to March 2023. Searches were performed using combinations of the following keywords: “pulsed field ablation”, “PFA”, “cryoballoon”, or “cryoablation”, together with “atrial fibrillation”. Only English-language articles were reviewed for eligibility. Identified studies were critically evaluated for inclusion in the final indirect comparison following best reporting practices from PRISMA guidelines [18], based on a qualitative assessment of study design, patient eligibility, patient baseline characteristics, and outcomes criteria.

Study Eligibility

Eligible studies had one of the following study designs: (1) FDA IDE (the Food and Drug Administration Investigation Device Exemption) studies; (2) first-in-human studies; or (3) registered clinical trials. Only the clinical trials that were registered on a clinical trial registry (i.e., clinicaltrials.gov) were considered. Studies had to involve adult patients with PAF who were drug-refractory to at least one class of AAD and received either PFA or CBA. Studies were excluded if they included patients with persistent AF, were not registered on a clinical trial registry, or were observational studies or conference abstracts. Studies were screened by an independent reviewer and checked for accuracy by a second reviewer.

Outcomes

The primary outcome was the composite of major procedure- or device-related adverse events reported at up to 30 days post-procedure. Major procedure- or device-related adverse events included cardiac tamponade or perforation, major vascular access complication or bleeding, myocardial infarction (MI), pericardial effusion, pericarditis, phrenic nerve paralysis, stroke, cerebrovascular accident (CVA), thromboembolism, transient ischemic attack (TIA), atrio-esophageal fistula, device- or procedure-related death, and pulmonary vein (PV) stenosis. Of note, PV stenosis and atrio-esophageal fistula were included in the primary outcome regardless of the timepoint. These events were identified as known potential risks across multiple clinical trials of AF ablation devices [12, 13, 15, 19,20,21,22,23,24], and as per guidance of the 2017 expert consensus statement from the Heart Rhythm Society (HRS) [25]. The secondary outcome was the composite of several prespecified adverse events reported at any timepoint.

Data Extraction and Feasibility Assessment

Data on study design, patient baseline characteristics, treatments, and safety outcomes were extracted into a standardized template by a single reviewer and checked for accuracy by a second reviewer. Cross-trial heterogeneity may contribute to bias within indirect treatment comparisons [26]. To ensure that studies were similarly sufficient for indirect comparison, a qualitative assessment of study design, eligibility criteria, patient baseline characteristics, treatments, and outcome definitions across the studies was conducted in accordance with best practices [27,28,29]. Studies that were deemed sufficiently similar based on clinical opinion were included in the analysis.

Statistical Analyses

Comparative safety between PFA and CBA was assessed through an unadjusted indirect treatment comparison and confirmed through single-arm meta-analyses. The proportion of patients experiencing an adverse event was calculated, and the results were presented using risk differences and risk ratios. Corresponding 95% confidence intervals (95% CI) were calculated using the Wald and the adjusted-log procedures, respectively. Sensitivity analyses were conducted by using other methods (i.e., Katz-log [30], Bailey [31], Noether [32], and Koopman [33]) to estimate the 95% CI.

Single-arm meta-analyses on PFA and CBA were performed separately to confirm the results from the indirect comparison. The overall weighted proportion of patients experiencing an adverse event within the eligible studies was estimated and presented in forest plots. Random effects model was used to account for potential heterogeneity between studies. Heterogeneity between study results was assessed using the τ2 statistic, and the percentage of total variation across studies was assessed using the inconsistency index, I2. Fixed effect meta-analyses were also conducted.

Analyses were conducted using the R packages DescTools [34] and meta [35] for single proportions. All analyses were conducted using the R software version 3.6.1.

Sensitivity Analyses

Sensitivity analyses were performed to assess the impact of cross-trial heterogeneity on the primary outcome. CBA studies with notable outlier values in the baseline characteristics were excluded one at a time from the primary analysis. PFA studies were not excluded to ensure a sufficient sample for analyses.

Statement of Ethics Compliance

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

Results

Study Inclusion and Feasibility Assessment

A total of 170 PFA and 158 CBA studies were identified as potentially relevant from literature searches (Fig. 1). After excluding the studies that did not meet the eligibility criteria, a total of three PFA studies representing five trials (InspIRE [12], IMPULSE, PEFCAT, PEFCAT II [13], and PULSED AF [15]) and six CBA studies representing six trials (Cryo versus RF [20], SUPIR [24], plusONE [19], STOP AF Post-Approval Study [21], FreezeAF [23], and Fire and Ice [22]) were deemed eligible for inclusion.

Fig. 1
figure 1

PRISMA flow diagram of the targeted literature review for A pulsed field ablation and B cryoballoon ablation. CBA cryoballoon ablation, PAF paroxysmal atrial fibrillation, PFA pulsed field ablation, T&A title and abstract

Full size image

All included studies were registered clinical trials (Table 1). All PFA trials were prospective, multi-center, single-arm trials with 12 months of follow-up. Four of the six CBA trials were prospective, randomized controlled trials with 12–18 months follow-up. The remaining two CBA trials were prospective, single-arm trials with 3.4 months (median) and 34.3 (mean) months of follow-up, respectively.

Table 1 Overview of key study characteristics in PFA and CBA clinical trials
Full size table

In total, the PFA studies consisted of 497 patients and the CBA studies consisted of 1113 patients. The distribution of baseline characteristics including age, gender, body mass index (BMI), left atrial diameter (LAD), and left ventricular ejection fraction (LVEF) were comparable across the studies (Table 2). The mean age of patients in the included studies ranged between 56 and 64 years. The distribution of male population was between 57.5% and 73.6% across the studies. A slightly lower male proportion (42.9%) was reported for the SUPIR study. The mean LAD values were 38.0–42.0 mm, and the mean LVEF was 57.9–64.2% across studies.

Table 2 Overview of patient baseline characteristics in PFA and CBA clinical trials
Full size table

Medical characteristics for the patients were also comparable among the PFA and CBA trials (Table 2). The ranges of proportion for different comorbidities were as follows: diabetes (4.8–16.0%); hypertension (34.6–61.5%); coronary artery disease (CAD; 0.7–20.7%), and stroke/TIA (2.1–9.0%). Overall, the included studies were deemed sufficiently similar by clinical opinion to conduct an unanchored indirect treatment comparison for safety outcomes.

Comparative Safety of PFA and CBA

Composite of Major Adverse Events

The indirect comparison revealed that PFA was associated with a statistically significantly lower risk of major adverse events than CBA, with a risk difference of  − 4.3% (95% CI  − 5.8,  − 2.8) and a risk ratio of 0.16 (95% CI 0.07, 0.45) (Table 3).

Table 3 Summary of indirect treatment comparison and single-arm meta-analyses of composite adverse events between PFA and CBA trials
Full size table

Results were confirmed in single-arm meta-analyses, which showed that 0.4% (95% CI 0.0, 1.3) of patients who underwent PFA experienced major adverse events, in contrast to 5.6% (95% CI 2.6, 8.6) of patients who underwent CBA (Table 3, Supplementary Fig. 1). Higher risk of major adverse events with CBA was primarily driven by phrenic nerve paralysis (PFA: 0 event vs. CBA: 33 events) and major vascular access complication or bleeding (PFA: 1 event vs. CBA: 12 events) (Table 4).

Table 4 Number of events for individual safety outcomes in pooled PFA and CBA trials
Full size table

Composite of Prespecified Adverse Events

The indirect comparison showed that PFA was associated with a statistically significantly lower risk of prespecified adverse events than CBA, with a risk difference of  − 2.5% (95% CI  − 4.4,  − 0.5) and a risk ratio of 0.53 (95% CI 0.31, 0.96) (Table 3).

Results were confirmed in single-arm meta-analyses, which showed that 2.7% (95% CI 1.2, 4.1) of patients who underwent PFA experienced an adverse event, in contrast to 5.8% (95% CI 2.7, 9.0) of patients who underwent CBA (Table 3, Supplementary Fig. 2).

Sensitivity Analyses

Four studies were identified with potential heterogeneity in baseline patient characteristics: SUPIR (duration of AF), plusONE (CAD and duration of AF), FreezeAF (hypertension), and Fire and Ice (hypertension) (Table 2). Exclusion of these studies one-at-a-time from the main analysis showed statistically significantly lower risk of major adverse events with PFA than with CBA (Supplementary Table 1).

Discussion

Pulsed field ablation (PFA) is emerging as an alternative to thermal-based catheter ablation technologies such as CBA, and multiple PFA systems have demonstrated their clinical effectiveness in single-arm clinical trials [12,13,14,15]. With limited head-to-head evidence comparing PFA to CBA available, an indirect treatment comparison of PFA and CBA was conducted to evaluate the comparative safety for PVI in patients with drug-refractory PAF. The present analysis shows that PFA has a significantly favorable safety profile compared to CBA. PFA was associated with 84% lower risk of major adverse events and 47% lower risk of prespecified adverse events than CBA. Results were driven by higher rates of major vascular access complication or bleeding, phrenic nerve injury, and pericardial effusion with CBA. Notably, there were no phrenic nerve injuries observed in the PFA trials included in the present study. Events of atrio-esophageal fistula or PV stenosis were also not observed in the included PFA trials. To the authors’ knowledge, this is the first indirect comparison of the safety profile between these two technologies using high-quality data from FDA IDE studies and registered clinical trials.

The evidence base supporting PFA in the treatment of drug-refractory PAF is evolving. ADVENT is the first randomized controlled trial to compare the effectiveness and safety of PFA (FARAPULSE; Boston Scientific, USA) (n = 301) or thermal ablation (n = 296) in adults with drug-refractory PAF [17]. Patients randomized to thermal ablation received either CBA (n = 132) or radio-frequency catheter ablation (n = 164), with each method planned to constitute approximately 50% of the aggregated control group [17]. The ADVENT trial showed that PFA was non-inferior to thermal ablation for the primary safety endpoint of acute- and chronic device-related and procedure-related serious adverse events (2.1% vs. 1.5% in thermal ablation, with posterior probability of non-inferiority > 0.999) [17]. In contrast, results from the current indirect comparison that pooled safety events from multiple PFA and CBA studies showed an improved safety profile with PFA relative to CBA, driven by phrenic nerve injury. Interestingly, no phrenic nerve injury was reported with thermal ablation in the ADVENT trial. Results from a real-world, retrospective, observational study comparing PFA (n = 200) and CBA (n = 200) also showed that PFA had an improved safety profile compared to CBA, with a numerically lower overall complication rate than CBA (3.0% vs. 6.5%; p = 0.1) that was driven by a higher rate of phrenic nerve injury with CBA (7.5% vs. 1.0%; p = 0.001) [36].

Overall, safety events across multiple novel PFA systems are reportedly low, with the rates of primary adverse events (assessed at 7 and 30 days after ablation) as low as zero for the VARIPULSE catheter and TRUPULSE generator system (Biosense Webster, USA) [12], 0.7% for PulseSelect (Medtronic, USA) [15], and up to 2.5% for FARAPULSE (Boston Scientific) [13]. Exceptionally low rates of major adverse events such as cardiac tamponade, pericardial effusion, and vascular complications have been reported with the PulseSelect and FARAPULSE systems, while no events were observed with the VARIPULSE/TRUPULSE system [12].

The current indirect comparison provides meaningful insight into the comparative safety profile of PFA and CBA. A strength of this analysis is that it only included patients with PAF who were drug-refractory to at least one AAD from a high-quality evidence base of registered clinical trials. Indirect treatment comparisons rely on the assumption that trials are sufficiently similar such that any underlying differences in patient populations would not bias the comparative effect estimate. A qualitative assessment of cross-trial heterogeneity was conducted a priori, the results of which showed sufficiently similar baseline characteristics across the trials. This was supported by clinical opinion. Nevertheless, this analysis was associated with limitations inherent to naïve indirect comparisons, including increased susceptibility to potential bias stemming from cross-trial heterogeneity. The impact of cross-trial heterogeneity on the analysis was explored in meta-analyses and sensitivity analyses. The results from meta-analyses using the random-effects model to account for cross-trial differences within PFA and CBA studies aligned with those from the indirect comparison. Additionally, the results for major adverse events were robust to excluding studies with outlier baseline patient characteristics one at a time in sensitivity analyses. However, cross-trial heterogeneity could not be fully explored within PFA studies owing to the need for a sufficient sample to conduct analyses. Anchored comparisons were not feasible due to limited randomized trials with a common comparator. Further, the ADVENT trial, which compared PFA to thermal ablation, was published outside the literature search window and therefore not considered for inclusion in this analysis.

Conclusion

Overall, the indirect treatment comparison and meta-analyses of the safety profile of PFA and CBA showed that PFA had significantly lower risks of major and prespecified procedure- or device-related adverse events than CBA, mainly driven by the higher rates of phrenic nerve palsy and vascular complications with CBA. These findings imply that PFA may be a reasonable alternative to CBA for PVI in patients with drug-refractory PAF.