Intravascular haemolysis following pulsed field ablation pulmonary vein isolation for atrial fibrillation: a systematic review

Abstract

Background

Pulsed field ablation (PFA) has emerged as a promising non-thermal alternative to conventional atrial fibrillation (AF) ablation techniques. However, intravascular haemolysis has been increasingly recognised as a potential complication, with variable incidence and clinical significance.

Objective

To systematically review the available clinical evidence on PFA-related haemolysis, focusing on biochemical markers, clinical manifestations and device-specific differences.

Methods

PubMed, Embase and Cochrane databases were searched until 20 May 2025 for clinical studies evaluating primarily PFA-pulmonary vein isolation for AF that reported haemolysis, acute kidney injury (AKI) or relevant biomarker changes. The primary outcome was evaluation of incidence and biochemical evidence of PFA-related haemolysis. Secondary outcomes included incidence of AKI and its clinical consequences.

Results

12 studies (≈20 000 patients) were included. Biomarker evidence of haemolysis was consistent, with postablation lactate dehydrogenase elevations of 250–438 U/L and bilirubin 15–48 µmol/L, often accompanied by reduced haptoglobin and elevated free haemoglobin. Incidence of haemolysis varied widely (0–94.3%), reflecting heterogeneity in definitions and reporting. Clinical sequelae were uncommon: haemoglobinuria was observed in five studies, and AKI occurred in 83 patients (0.4%), 12 requiring transient dialysis. All returned to baseline renal function except one patient with severe chronic kidney disease. Procedural factors and catheter design may influence haemolysis burden. Observations of lower haemolytic biomarker changes with devices such as PulseSelect, Affera and Volt are preliminary and require confirmation, given the predominance of Farawave data.

Conclusions

Haemolysis is a reproducible biochemical outcome of PFA, but clinically significant events such as AKI are rare and usually reversible. Catheter design, energy delivery and patient baseline renal function are likely to modulate haemolysis risk. Standardised haemolysis definitions and prospective head-to-head comparisons across PFA platforms are needed to clarify clinical relevance and optimise safety.

PROSPERO registration number

CRD420251069612.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Pulsed field ablation (PFA) is a novel non-thermal technology for atrial fibrillation with a favourable safety profile, but reports of haemolysis and occasional acute kidney injury have raised concerns.

WHAT THIS STUDY ADDS

• This systematic review of ~20 000 patients shows that haemolysis is a consistent biochemical finding after PFA but rarely leads to clinical complications, with risk potentially influenced by application burden, catheter design and baseline renal function.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings support careful procedural planning, periprocedural hydration and closer monitoring in patients with renal impairment, while emphasising the need for regular preablation and postablation biochemical testing and comparative evaluation of PFA systems.

Introduction

Atrial fibrillation (AF) is the most prevalent cardiac arrhythmia and can be associated with significant morbidity with an increased risk of thromboembolic events and heart failure.^{1 2} Recent evidence increasingly supports catheter ablation as a first-line strategy for rhythm control in patients with paroxysmal AF, whereas its superiority over drug therapy in persistent AF remains less clear.³ Nonetheless, the rising adoption of catheter ablation has spurred the development of innovative technologies. Among these, pulsed field ablation (PFA) has emerged as a promising non-thermal alternative to conventional thermal ablation modalities such as radiofrequency ablation (RFA) and cryoballoon ablation.⁴ PFA uses high-amplitude electrical pulses to induce irreversible electroporation, resulting in cell death by creating permanent nanopores within the cell membranes.⁴ This technique can selectively ablate myocardial tissue due to its different susceptibility to electric fields compared with surrounding non-cardiac structures.⁵

This selective effect on myocardial tissue contributes to a favourable safety profile, with preclinical and early clinical studies demonstrating minimal injury to adjacent structures such as blood vessels, nerves and the oesophagus.^{6 7} However, emerging data have raised concerns regarding PFA-related intravascular haemolysis.⁸ The high-energy electrical fields used in PFA may disrupt red blood cell (RBC) membranes, leading to haemolysis and the release of free haemoglobin (fHb) and other intracellular contents. Biomarker evidence of haemolysis, specifically increases in lactate dehydrogenase (LDH), total bilirubin, fHb and corresponding reduction in haptoglobin, has been reported following PFA procedures.^{7 9 10} In some cases, this has been associated with haemoglobinuria and rarely subsequent acute kidney injury (AKI).^{9 11} These findings are supported by in vitro studies demonstrating significant RBC lysis and cardiomyocyte death with increased pulsed field energy.¹²

Despite these observations, the clinical significance of PFA-related haemolysis remains unclear. Some studies have reported transient, asymptomatic biomarker changes, while others suggest higher haemolysis rates with end-organ injury.^{7 13 14} More importantly, consensus regarding the definitions of clinical or biochemical haemolysis has remained non-standardised across the literature, resulting in inconsistent findings.

Prior reviews, including Popa et al and Xu et al, have provided important insights into the mechanisms, risk factors and potential prevention strategies for PFA-related haemolysis.^{7 8} Our review builds on these by applying a systematic Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)-guided methodology, conducting a formal thorough search of the literature, providing a formal risk-of-bias assessment and incorporating four additional studies published in 2025 evaluating newer PFA platforms. By doing so, we expand the evidence base to more than 20 000 patients and provide a structured appraisal of the incidence, biochemical changes and clinical relevance of PFA-related haemolysis. Given the growing adoption of PFA in clinical practice, a comprehensive and methodologically rigorous synthesis is timely to better inform clinicians and researchers.

Methods

This systematic review is reported according to the PRISMA guidelines.¹⁵ This review is registered with PROSPERO (CRD420251069612).

Search strategy and selection criteria

The population, intervention, comparator, outcome framework was used to develop the research question and inclusion criteria. The population comprised patients with AF. The intervention was PFA. The comparator group was not applicable as this was a single-arm study; however, some studies did include comparisons to RFA. The outcome was dependent on each study and included the incidence of haemolysis or AKI and any biomarkers that were reported (LDH, total bilirubin, haptoglobin or fHb). Eligible studies included human clinical trials or observational studies of PFA for AF that reported haemolysis, AKI or at least one relevant biomarker. Publications were excluded if they were in vitro studies and conference abstracts, lacked haemolysis-related data or did not involve PFA. While some eligible studies included subsets of patients who underwent additional posterior wall isolation (PWI) or mitral isthmus ablation in addition to pulmonary vein isolation (PVI), studies in which PWI (with or without PVI) was the primary ablation strategy were also excluded. PubMed (incorporating MEDLINE), Ovid Embase and Cochrane were searched from respective database inception to 20 May 2025 for studies of any design and in any setting. No limitation to language or publication restrictions was implemented. The full search strategies can be found in online supplemental appendix 1.

Data extraction and bias assessment

Titles and abstracts were independently screened by two reviewers (TK and CT). A web application (Rayyan, Qatar Computing Research Institute, Ar-Rayyan, Qatar) was used to facilitate blinded screening of titles and abstracts.¹⁶ Disagreements were resolved by consensus. Data from the selected studies were then extracted by both reviewers together using a standardised data collection table. Extracted information included incidence of haemolysis, AKI and biochemical data following procedure.

Results

Search results

After eliminating duplicate records, the literature search identified 83 unique studies. Of these, 12 studies met the inclusion criteria for this systematic review (refer to figure 1). A list of the studies that were excluded at the stage of full-text review, with justification of exclusion for each potentially relevant study, is outlined in detail in online supplemental appendix 2.

Figure 1. PRISMA flow chart. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Of the 12 studies extracted, four were pre-post studies without control groups, five were prospective cohort studies, two were retrospective observational studies and one was a prospective controlled trial. No randomised controlled trials were identified. The corresponding National Institutes of Health (NIH) Quality Assessment Tool was applied to each study according to its design.¹⁷ Two independent reviewers (CT and TK) conducted quality assessments for all articles.

Study characteristics

The 12 studies were published between 2024 and 2025 reflecting the novelty of PFA. They were conducted across Europe (n=6), Asia (n=1), North America (n=4) and in multinational cohorts (n=1). The studies encompassed a total of approximately 20 000 patients with sample sizes ranging from 47 to 17 642. All studies examined patients undergoing PFA for AF primarily for PVI, though several also included subsets of patients with additional target lesions warranting posterior wall and/or mitral isthmus ablation.

Catheter types varied, with the Farawave pentaspline catheter being the most common. The largest of the studies, MANIFEST-17K, contributed 17 642 patients (≈87.7%), almost all treated with the Farawave pentaspline catheter. Conversely, data from other platforms (Volt, Affera, PulseSelect) were limited to small single-centre experiences.1118,20 The number of ablation applications also varied widely depending on extended ablation beyond PVI, with maximum reported being over 200.

LDH and total bilirubin were the most common biomarkers reported. However, there was marked heterogeneity in the biomarker reporting, diagnostic thresholds, timing of postprocedural measurements and the lack of paired baseline data. These factors precluded a valid pooled quantitative meta-analysis. We therefore performed a structured narrative synthesis to summarise trends and associations across the included studies. Where paired preprocedural and postprocedural values were made available, we present study-level comparisons in a forest-style format for transparency, but without pooled estimates (figures24). A summary of study characteristics and biomarker values is also presented in tables1 2, respectively. The included studies varied widely in risk of bias on critical appraisal with the NIH Quality Assessment Tool; however, most studies were considered moderate risk of bias (online supplemental appendix 3).

Figure 2. Summary of LDH (U/L) before and after ablation across included studies with available data. IV, inverse variance; LDH, lactate dehydrogenase.

Figure 4. Summary of haptoglobin (g/L) before and after ablation across included studies with available data. IV, inverse variance.

Table 1. Summary of the 12 included studies evaluating haemolysis following PFA for AF.

Study	Country	Study population	Types of PFA catheters	Definition of haemolysis	Incidence of haemolysis	Incidence of AKI
Auf der Heiden et al13	Germany	150 patients	Farawave	Bilirubin ↑ LDH ↑	Not reported	0%
Lo et al19	USA	392 patients	Volt	Bilirubin ↑ LDH ↑ Haptoglobin ↓	No clinically significant haemolysis reported	0%
Xuan et al20	China	94 patients	CardiPulse	Bilirubin ↑	30/94 (31.9%)	0%
Cho et al23	USA	Not reported	Farawave and PulseSelect	No definition for haemolysis given	14 patients	9 patients
Stojadinović et al10	Czech Republic	60 patients	Farawave	LDH ↑ fHb ↑	29/60 (48.3%)	0%
Popa et al7	France	145 patients	Farawave	Bilirubin ↑ LDH ↑ Haptoglobin ↓ fHb ↑ Spectrophotometry	33/35 (94.3%) via spectrophotometric analysis 94/98 (95.9%) had haptoglobin ratio <0.85 and fHb >200	4/124 (3.2%)
Ekanem et al11	Multinational	17 642 patients	Farawave	Haemolysis-related renal failure (oliguria/anuria)	6/17 642 (0.03%)	5/17 642 (0.03%)
Osmancik et al21	Czech Republic	47 patients	Farawave	LDH ↑ Bilirubin ↑ Haptoglobin ↓ RBC microparticles ↑	Not reported	0%
Mohanty et al14	USA	103 patients (28 no hydration group+75 control)	Farawave	LDH ↑ Haptoglobin ↓ Haemoglobinuria	No hydration group: 21/28 (75%) Control: 0/75 (0%)	No hydration: 4/28 (14.3%) Control: 0/75 (0%)
Venier et al9	France	68 patients	Farawave	Haptoglobin ↓ Haemoglobinuria	19/68 (28%)	2/68 (2.94%)
Lakkireddy et al18	USA	773 patients	Farawave, PulseSelect, Affera, Varipulse	LDH ↑ Haptoglobin ↓ fHb ↑	Not reported	51/773 (6.7%)
De Smet et al22	Belgium	152 patients	Farawave, Varipulse	Haptoglobin ↓ Haemoglobinuria and AKI	Not reported	8/152 (5.26%)

AF, atrial fibrillation; AKI, acute kidney injury; fHb, free haemoglobin; LDH, lactate dehydrogenase; PFA, pulsed field ablation; RBC, red blood cell.

Table 2. Reported postprocedural biomarker levels across the included studies.

Study	LDH (U/L) (baseline 150–200)*	Total bilirubin (µmol/L) (baseline 9–16)*	Haptoglobin (g/L) (baseline >1.2)*	fHb (mg/L) (baseline <50)*	Other
Auf der Heiden et al13	Increased (no numerical value reported)	Increased (no numerical value reported)	Nil	Nil	Nil
Lo et al19	254.9±60.0	17.1±10.3	1.0±0.4	Nil	Nil
Xuan et al20	337.3±47.2	47.8±12.6	Nil	Nil	Nil
Cho et al23	Nil	Nil	Nil	Nil	Nil
Stojadinović et al10	394.0±132.2	Nil	Nil	556.0±395.7	Nil
Popa et al7	352.7±115.7	21.3±11.3	0.50±0.40	592.8±330.6	Nil
Ekanem et al11	Nil	Nil	Nil	Nil	Nil
Osmancik et al21	283.6±64.7	Nil	0.47±0.35	Nil	Peak RBC microparticles 12-fold higher than preablation (924.2 RBCµ/µL)
Mohanty et al (group 1)14	438.2±56.7	29.1±19.8	0.10±0.08	Nil	Nil
Mohanty et al (group 2)14	Nil	15.4±9.8	Nil	Nil	Nil
Venier et al9	376.7±101.5	26.2±13.6	0.24±0.36	Nil	Nil
Lakkireddy et al18	Nil	Nil	Nil	612±125	Only delta change reported; no absolute postprocedural levels
De Smet et al22	250.4±57.2†	21.5±12.7†	0.71±0.48†	456±347†	Nil

Studies that reported medians and IQRs were converted to mean±SD using the method of Wan et al for consistency.28

^*Baseline ranges were derived from studies that reported preablation values.

^†Weighted average was calculated for this study due to style of reporting.

fHb, free haemoglobin; LDH, lactate dehydrogenase; RBC, red blood cell.

Figure 3. Summary of total bilirubin (g/L) before and after ablation across included studies with available data. IV, inverse variance.

Incidence and biomarker trends of haemolysis

The reported incidence of haemolysis varied widely across studies, ranging from 0% to 94.3%. This variability reflects the lack of standardisation in haemolysis definitions and variety of diagnostic criteria applied. Two studies inferred haemolysis solely based on postprocedural elevations in LDH and bilirubin.^{13 20} Seven additional studies complemented this with the inclusion of haptoglobin measures for a more robust haemolysis profile.^{7 9 14 18 19 21 22} Four studies evaluated a more direct result of haemolysis by reporting fHb, while one study uniquely employed flow cytometry to directly quantify RBC microparticles as a measure of direct haemolysis.^{7 10 18 21 22} Finally, one other study relied solely on clinical manifestations in haemoglobinuria, oliguria and AKI without reporting biomarker correlation, and another study did not disclose the criteria used to diagnose haemolysis.^{11 23}

Where reported, postablation LDH levels ranged from 250.4±57.2 U/L to 438.2±56.7 U/L, while total bilirubin ranged from 15.4±9.8 µmol/L to 47.8±12.6 µmol/L. fHb levels were markedly elevated in the four studies that reported it, peaking at 612±125 mg/L with haptoglobin levels decreasing as low as 0.10±0.08 g/L. ^{7 10 14 18} Notably, Lo et al’s study was the only one that reported no significant change in haptoglobin from baseline with a postablation haptoglobin of 1.0±0.4 g/L, contrasting the substantial reduction seen in other studies.¹⁹ One study performed spectrophotometric analysis of postablation samples and reported a 12-fold increase in RBC microparticles, peaking at 924.2 RBCµ/µL.²¹

Importantly, not all studies included preablation baseline values and thus precluded consistent assessment of biomarker change following PFA. However, to further illustrate biomarker dynamics, figures24 demonstrate pooled paired preablation and postablation measurements of LDH, total bilirubin and haptoglobin across studies that reported these outcomes. Consistent postprocedural rises in LDH and bilirubin, alongside reductions in haptoglobin, are evident. Error bars indicate SDs. For studies lacking baseline values, reference intervals from the Royal College of Pathologists of Australasia (RCPA) were applied. Normal reference intervals defined by the RCPA were used.

Haemoglobinuria and AKI

Clinical consequences from PFA-related haemolysis were infrequent and inconsistent across the analysed population. Haemoglobinuria was documented in five studies, with incidence rates reaching as high as 75%.^{7 9 11 14 22} AKI was reported in seven studies with incidence ranging from 0.03% to 14.3%.^{7 9 11 14 18 22 23} Although the majority of AKI cases were transient and resolved either spontaneously or with supportive measures, three studies described a total of 11 cases where temporary haemodialysis was required.^{11 18 23} A cumulative total of 83 AKI events were reported among 20 118 patients (0.4%), with 12 of the 83 requiring dialysis.^{7 9 11 14 18 22 23} All patients who reported AKI returned to baseline renal function, except for one who had severe chronic kidney disease (CKD) with an estimated glomerular filtration rate of 14.²²

Discussion

This systematic review provides a comprehensive and updated synthesis of the current evidence on haemolysis following PFA for AF, focusing on biomarker trends and clinical significance. Intravascular haemolysis appears to be a predictable biochemical outcome of PFA, evidenced by consistent increases in LDH, bilirubin and associated decreases in haptoglobin levels across most studies. However, the translation into clinical complications is uncommon. Although 11 of the 12 included studies reported biomarker changes consistent with haemolysis, only 83 reported clinical manifestations, including haemoglobinuria or AKI.79,11 13 14 18 21 23 Among nearly 20 000 patients, the overall incidence of AKI was low (0.4%), and no study reported symptomatic anaemia or jaundice attributable to haemolysis. Notably, hydration status emerged as a modifiable risk factor. In a comparative cohort, Mohanty et al demonstrated that the implementation of routine postablation hydration reduced AKI incidence from 14.3% to 0%, without any reported cases of fluid overload or pulmonary oedema.¹⁴ Multivariable analysis showed volume of fluid supplementation to be an independent predictor of AKI. While these observational data are encouraging, randomised evidence is lacking. We therefore recommend routine systematic use and reporting of periprocedural hydration practices in future PFA studies to better define their benefit and safety. Furthermore, baseline renal function may increase susceptibility to clinically significant haemolysis. De Smet et al showed that AKI predominantly affected patients who were elderly and had pre-existing CKD and diabetes mellitus.²² These findings suggest that while biochemical haemolysis is a consistent feature of PFA, its clinical consequences are rare and influenced by patient-specific risk factors which may be mitigated by preventative strategies such as periprocedural hydration.

Procedural factors

Preclinical studies suggest that the severity and rates of intravascular haemolysis following PFA appear to be influenced by several procedural factors, including the number of applications delivered, quality of catheter–tissue contact and energy delivery.24,26 While haemolysis has been observed across a range of application counts, the data suggest that biochemical evidence of haemolysis becomes more pronounced once application numbers exceed 54–74.^{7 10} Clinically significant AKI, although reported rarely, generally occurs only beyond 90–100 applications.^{9 11} Despite manufacturer recommendations to limit ablation to eight applications per vein (32 total),¹³ this threshold is often exceeded in clinical practice due to incomplete isolation or the need for additional ablation lines in the posterior wall, mitral isthmus or cavotricuspid isthmus. In the large MANIFEST-17K registry, for instance, all five patients who developed an AKI received an average of 143 applications and all received further ablation lines beyond standard PVI.¹¹ Fiserova et al have shown a clear dose–response relationship between electric field strength and haemolysis, with significant increases in fHb observed at intensities ≥1000 V/cm, and near-complete RBC lysis occurring above 2500 V/cm in preclinical studies.¹² In clinical settings, PFA systems typically operate at voltages between 1800 and 2200 V, but the true electric field intensity at the catheter–tissue interface is not disclosed by manufacturers and may exceed thresholds.

Catheter design and intersystem comparison

Catheter–tissue interface dynamics also plays an important role in modulating haemolytic burden during PFA. Suboptimal contact between catheter and the endocardial surface can lead to uneven energy distribution and greater dissipation of electrical fields into the circulating blood pool, thereby increasing the likelihood of RBC membrane disruption.²⁵ While this mechanism has been discussed in preclinical models, clinical studies have not routinely reported catheter contact quality, likely due to limitations in real-time contact assessment. Catheters with real-time contact sensing, such as the Volt balloon-in-basket system, can deactivate poorly apposed splines and focus energy delivery, potentially reducing unnecessary blood pool exposure.¹⁹ By contrast, Farawave pentaspline-based systems, which account for the majority of clinical haemolysis data to date, do not offer such contact-sensing capabilities. This, combined with the significantly fewer applications required for PVI with Volt compared with the Farawave pentaspline system, may explain why no cases of clinically significant haemolysis have been reported in the single clinical review using the Volt system, though the data are limited and should be interpreted as hypothesis generating.

Furthermore, comparative studies suggest that catheter shape also influences haemolysis risk. The circular and spherical design of the PulseSelect and Affera could potentially avoid RBC destruction due to targeted delivery of energy. This finding is supported by findings in small comparative studies from Cho et al and Lakkireddy et al, who demonstrated significantly decreased haemolytic biomarker change when evaluating the PulseSelect and Affera systems compared with Farawave.^{18 23} Possible explanations for this include the lower voltage used by such systems (1500 V vs 2000 V) and improved contact profiles with avoidance of strong, central field generation.²⁷ These preliminary findings are intriguing but remain exploratory given sample size limitations and lack of head-to-head randomisation. However, it underscores the importance of both catheter architecture and energy delivery strategy in mitigating haemolysis and highlights the need for further head-to-head evaluation across platforms to validate these preliminary observations.

Limitations

This study has limitations. A major challenge was the significant heterogeneity across included studies, particularly in terms of biomarkers used to define haemolysis, biomarkers reported, PFA catheter systems and adjunctive lesion sets applied beyond PVI. This variability limited the ability to conduct quantitative synthesis or meta-analysis. Instead, a narrative synthesis approach was employed to qualitatively summarise trends. The main limitation in evaluating PFA-related haemolysis across multiple studies is the lack of standardised definition of haemolysis or diagnostic threshold. Biomarker cut-offs used to define haemolysis, whether based on LDH, bilirubin, fHb or haptoglobin, varied widely and, in several cases, were not clearly defined. The absence of baseline biomarker values in several studies further obscures the magnitude of change attributable to PFA. For example, the MANIFEST-17K trial, despite including over 17 000 patients, did not report any biochemical markers of haemolysis, instead relying on clinical diagnosis of AKI as a clinical sequela of haemolysis.¹¹ Thus, the haemolysis incidence ranging from 0% to 94.3% could reflect differences in definitions rather than true biological variability and should be interpreted with caution. It is important to note that most included studies were observational in nature and were deemed of moderate level of bias. Moreover, in the absence of systematic biomarker testing of all participants, subclinical or transient haemolysis events may have been under-reported, particularly in retrospective cohorts. Until a uniform, validated definition of biochemical haemolysis is adopted, cross-study comparisons of haemolysis rates remain limited and place the overall confidence in the current evidence at low to moderate at best.

Furthermore, the generalisability of findings is limited by the predominance of a single ablation system. The majority of included studies used the Farawave pentaspline catheter. Given potential differences in energy delivery schemes, electrode configuration and catheter–tissue interaction across platforms, it remains unclear whether haemolysis profiles observed with Farawave are representative of other devices. This imbalance limits interdevice comparison and prevents a comprehensive evaluation of haemolysis risk across the growing landscape of PFA technologies.

Conclusion

This systematic review highlights that while intravascular haemolysis is a reproducible biochemical consequence of PFA, clinically significant sequelae such as AKI appear to be rare and typically transient in the available data (figure 5). The severity of haemolysis is hypothesised to correlate with procedural factors and catheter design. A standardised definition of haemolysis and broader evaluation across diverse PFA platforms are essential to improve our understanding of its clinical relevance and to optimise procedural safety. Given the observational nature of the evidence and the predominance of a single system, the findings in this study remain exploratory and require confirmation in larger scale prospective studies evaluating different PFA delivery systems and catheters.

Figure 5. Central graphical abstract. AKI, acute kidney injury; LDH, lactate dehydrogenase; NIH, National Institutes of Health; PFA, pulsed field ablation.

Data availability statement

Data are available in a public, open access repository.