Cryoablation vs. radiofrequency ablation for treatment of paroxysmal atrial fibrillation: a systematic review and meta-analysis

Transcription

1 Europace (2017) 19, doi: /europace/euw330 CLINICAL RESEARCH Ablation for atrial fibrillation Cryoablation vs. radiofrequency ablation for treatment of paroxysmal atrial fibrillation: a systematic review and meta-analysis Yi-He Chen 1, Zhao-Yang Lu 1, Yin-Xiang 1, Jian-Wen Hou 1, Qian Wang 1, Hui Lin 2 *, and Yi-Gang Li 1 * 1 Department of Cardiology, Xinhua Hospital Affiliated to The Medical School of Shanghai Jiaotong University, 1665 Kongjiang Road, Shanghai , China; and 2 Department of Respiratory, The Second Affiliated Hospital of Wenzhou Medical University, Wenzhou , China Received 18 August 2016; accepted after revision 19 September 2016; online publish-ahead-of-print 8 January 2017 Aims Cryoablation is a promising alternative technique to RF ablation for treating paroxysmal AF with encouraging results. However, data about the efficacy and safety comparison between cryoablation and RF ablation is still lacking.... Methods We systematically search the PubMed, the Cochrane Library, MEDLINE and Google Scholar databases, and finally and results identify 16 eligible studies including 7195 patients (2863 for cryoablation; 4332 for RF ablation). Freedom from AF/ atrial tachycardial replase is slightly higher in cryoablation than RF ablation during a median 12 months of followup, with no statistical significant (RR: 1.05, 95% CI: , P ¼ 0.159). In cryoablation, the procedure time is substantially shortened (WMD: 27.66, 95% CI: to 10.08, P ¼ 0.002), whereas the fluoroscopy time is identical to RF ablation (WMD: 0.37, 95% CI: 2.78 to 2.04, P ¼ 0.763). Procedure-related adverse events in cryoablation are parallel with that in RF ablation (RR: 1.08, 95% CI: , P ¼ 0.159).... Conclusions Compared with RF ablation, cryoablation present a comparable long-term AF/atrial tachycardial-free survival and procedure-related adverse events. Meanwhile, cryoablation markedly shorten the procedure time, nonetheless, with negligible impact on the fluoroscopy time. Keywords Cryoablation Radiofrequency ablation Paroxysmal atrial fibrillation Meta-analysis Introduction Catheter ablation is now the well established interventional approach for treating symptomatic drug-refractory, paroxysmal AF, with a Class I level A recommendation in ESC and ACC/AHA guidelines. 1,2 Electrical isolation achieved by creating contiguous and transmual lesion between pulmonary veins and atrium has become the cornerstone ablation strategy for paroxysmal AF. 3 RF current is the prevailing energy source of catheter ablation worldwide, nonetheless, long-term atrial arrhythmia-free survival is still unsatisfactory. 4 Moreover, it has also been challenged by the emerging novel and alternative techniques, especially the cryoablation, due to the complexity of manipulation and potential severe complications (i.e. pulmonary vein stenosis, cardiac perforation, esophageal fistulas and major bleeding). 3,5,6 In contrast to the point-by-point focal RF energy delivery, catheter ablation using cryothermal energy simplifies the procedure for creating circumferential transmural lesions through balloon-based technology, which offers a homogeneous freezing area and thus results in irreversible damage of myocardium in a single application. 7 Likewise, cryoablation is associated with a shorter learning curve and may be less influenced by individual difference in experience and aptitude. 8,9 Therefore, it provides an invaluable advantage of achieving more reproductive and optimal results. 10 Recently, cryoablation is increasingly used for treating AF in clinical practice. Data from previous studies have demonstrated the cryoablation as a highly effective alternative approach in treatment of paroxysmal AF, with non-inferior outcomes in comparison with RF ablation However, the relatively small sample size, material advance (i.e. CF-sensing catheter or second generation cryoballoon) and improvement of ablation * Corresponding author. Tel: þ ; fax: þ address: drliyigang@outlook.com; linhui1203@126.com Published on behalf of the European Society of Cardiology. All rights reserved. VC The Author For Permissions, please journals.permissions@oup.com.

2 Cryoablation vs. radiofrequency ablation for paroxysmal AF 785 What s new? Cryoablation present a comparable long-term AF/atrial tachycardial-free survival compared with RF ablation. Cryoablation markedly shorten the procedure time, with negligible impact on the fluoroscopy time. Cryoablation has no impact on procedure-related adverse events. techniques, along with unassessed confounding variables may limit the strength and reliance of evidence from these studies. 13,14 Furthermore, data on the two remarkable RCTs comparing RF ablation with cryoablation (the FIRE AND ICE trial and the FREEZE trail) are currently available online, which provide a more comprehensive grasp of the cryothermal energy application in paroxysmal AF. 15,16 Therefore, we conduct this systemic literature review and metaanalysis to evaluate the efficacy and safety of cryoablation vs. RF ablation for treating paroxysmal AF in the context of expands in clinical evidence and the noticeable effect of cryoablation. Methods Literature strategy We systematically reviewed the relevant literature, comparing cryoablationvs.rfablationinpatientswithparoxysmalaf,bysearchingpubmed, the Cochrane Library, MEDLINE and Google Scholar databases from inception to July 2016, without any language restriction. The following keywords were used for search: atrial fibrillation or AF, AND cryoablation or cryothermal or cryoballoon or cryotherapy or cryo*, AND ablation or PVI. In addition, reference of review articles, meeting abstracts, letter, editorials as well as previous meta-analysis were also searched. Selection criteria We included all eligible studies that met the following inclusion criteria: (i) RCTs or observational studies; (ii) only paroxysmal AF; (iii) cryoablation vs. RF ablation; (iv) a minimum of 12 months of follow-up. For each included study, data of the following efficacy and safety endpoint were used and evaluated: AF/atrial tachycardia replase after a blanking period of 3 months, total procedure time, fluoroscopy time and procedurerelated adverse events. Studies that enrolled patients with an initial failed atrial ablation or surgery were excluded. Data extraction All data from included studies were independently extracted and assessed for further analysis by two reviewers (Zhao-Yang Lu and Yin Xiang). Any discrepancies were resolved through referral to a third reviewer. From each studies, information about the baseline patient characteristics (age, gender, left atrium diameter, LVEF, history of paroxysmal AF and comorbidities), study design, type of catheter, follow-up duration and the ablation characteristics of each energy type during the procedure, along with the pre-specified clinical endpoint (AF/atrial tachycardia replase, procedure time, fluoroscopy time and procedure-related adverse events) was extracted and tabulated. AF/atrial tachycardia replase was defined as documented episode of AF, atrial flutter or atrial tachycardia lasting for >30 seconds after the blanking period. Procedure-related adverse events were identified as occurrence of death, stroke or transient ischemic attack, phrenic nerve palsy, pericardial complications (effusion, tamponade or pericarditis), vascular complications (hematoma, pseudoaneurysm or arteriovenous fistula), bleeding and the other complications during the catheter ablation. Data synthesis and analysis Whenever possible, intention-to-treat data were used for evaluating clinical endpoints from included studies. The RRs and respective 95% CIs were computed for categorical variables (AF/atrial tachycardia-free survival and procedure-related adverse events) and, as appropriate, the WMD and respective 95% CIs were evaluated for continuous variables (procedure time and fluoroscopy time). We used random-effects models in this meta-analysis with the DerSimonian & Laird method. The extent of heterogeneity between studies was quantified with I 2 statistic and further estimated with the Cochran Q test. Publication bias was primarily assessed by visually studying funnel plot and further tested with Begg s and the Egger s test. Sensitivity analysis was performed to identify the influence of individual study on pooled estimates by removing one study each time. To explore the impact of potential effect modifiers on outcomes and the possible sources of heterogeneity, exploratory univariate metaregressions were performed. We conducted subgroup analysis according to the study design (RCT or observational study), study size (larger [>100 patients in each study] or smaller), use of CF-sensing catheter, generation of cryoballoon and follow-up duration (>12 months or 12 months). All data were analysed using SPSS version 20.0 (SPSS Inc., IL), and STATA software version 12.0 (StataCorp, College Station, Texas), with P values 0.05 considered statistically significant. Results Characteristics of included studies Thedetailsoftheliteraturesearchwereshowedintheflowdiagram (Figure 1). Our preliminary search identified 273 potentially articles using the aforementioned keywords. Of these,16 studies finally met the inclusion criteria after rigorous scrutiny. A comprehensive list of eligible studies was presented in Table 1, while Table 2 summarized and compared the baseline patients demographic characteristics in cryoablation and RF ablation from these studies. Of the 7195 patients, 2863 underwent cryoablation and 4332 underwent RF ablation. The mean age of patients was 59.9 years, with a history of paroxysmal AF ranged from 2.1 to 5.2 years. There were four RCTs, three retrospective cohort studies, eight perspective cohort studies and one ambidirectional cohort study (either prospectively or retrospectively). Cryoablation of pulmonary vein was performed with the second generation cryoballoon (Arctic Front AdvanceTM, Medtronic) in the majority of studies, however, there were still six studies utilized the previous version exclusively (Arctic FrontTM, Medtronic). For RF ablation, CF-sensing catheter was applied in five studies with an optimal CF 10 g during the procedure, nonetheless, no more than the upper limit of 50 g. In addition, almost all RF-mediated atrial electrical isolation was carried out with either a 3.5- or a 4-mm irrigated tip. Complete PVI was confirmed using a circular mapping catheter 20 or 30 min after last ablation. Additional adenosine test was performed in the study by Squara et al. 17 The mean follow-up duration ranged from 12 to 28 months. Furthermore, all studies were conducted in western country (four in Germany, three in France, one in Spain, one in Hungary, one in USA, twoinbelgium,twoinswitzerland,twoinuk).therewasnoevidence of significant differences for baseline patients demographic

3 786 Y.-H. Chen et al. Literature search from Pubmed, MEDLINE, the Cochrane Library and Google Scholar databases 600 articles after duplicates removed 600 articles reviewed 516 excluded based on title and abstract 84 full-text articles assessed for eligibility 68 articles excluded 35 No comparison between cryoablation and RF abaltion 29 Case, review, comments or editorial 1 History of atrial ablation 3 No endpoint of AF/atrial tachycardia-free survival 16 studies included in our analysis Figure 1 Flow chart of literature search. characteristics between cryoablation and RF ablation in spite of a tend toward less stroke and/or TIA in the former (Table 2). AF/atrial tachycardia-free survival The overall freedom from AF/atrial tachycardia replase was 65.6% (1823 of 2777 patients) with cryoablation and 60.1% (2555 of 4251 patients) with RF ablation during a median follow-up of 12 months. The pooled estimate of RR indicated that PVI achieved by cryoablation and RF ablation led to comparable long-term AF/atrial tachycardia-free survival during the follow-up (RR: 1.05, 95% CI: , P ¼ 0.159) (Figure 2). There was evidence of high heterogeneity (Cochrane Q P ¼ ; I 2 ¼72.5%). To explore the sources of heterogeneity, we performed meta-regression analysis to unmask the impact of study-level covariates (i.e. number of patients, age, gender, follow-up duration, use of CF-sensing catheter or generation of cryoballoon) on the relationship between the sources of energy during the catheter ablation and the AF/atrial tachycardia-free survival. However, there were no statistically significant interactions between the pooled RR and pre-specified covariates (P > 0.05) (Table 3). Furthermore, although systemic remove of each study did not influence the pooled estimates and respective P value, intriguingly, the heterogeneity was reduced to 15.4% by excluding the study of Providencia et al. (P ¼ 0.282). 10 The funnel plot appeared symmetrical by visual inspection (Figure 3), and further confirmed using the Begger s test (P ¼ 0.471) and Egger s test (P ¼ 0.297), which suggested no evidence of potential publication bias. Subgroup analysis was performed according to the study design, sample size, use of CFsensing catheter, follow-up duration as well as the generation of cryoballoon, detailed results were summarized in Table 4. Notably, no statistically significant change of the pooled estimates was observed among the subgroups. Procedure time and fluoroscopy time Procedure time and fluoroscopy time data were available for analysis in 15 studies representing 4889 patients (2256 for cryoablation, 2633 for RF ablation). During the ablation, cryoablation was associated with a shorter procedure time when compared with radiofrequency ablation (WMD: 27.66, 95% CI: to 10.08, P ¼ 0.002) (Figure 4A). To our surprise, fluoroscopy time in cryoablation was similar to that in RF ablation (WMD: 0.37, 95% CI: 2.78 to 2.04, P ¼ 0.763) (Figure 4B). There was evidence of significant heterogeneity for procedure time (I 2 ¼97.7, P ¼ 0.000) and fluoroscopy time (I 2 ¼89.4, P ¼ 0.000), respectively. Meta-regression failed to identify any significant interactions between the pooled estimates and prespecified covariates (Table 3). Likewise, sensitivity analysis by exclusion of each study also did not change the effect size and the P value. Procedure-related adverse events Data on procedure-related adverse events were available for metaanalysis in 15 studies. Events were occurred in 223 of 2759 patients (8.08%) with cryoablation and 333 of 4130 patients (8.06%) with radiofrequency ablation. The pooled estimates showed no significant difference between these two approaches for the ablation procedure-related adverse events (RR: 1.08, 95% CI: ,

4 Cryoablation vs. radiofrequency ablation for paroxysmal AF 787 Table 1 Characteristics of the studies included in meta-analysis Author Year State Design Patients... Age Gender AF history Follow-up Catheters... Ablation... Cryo RF (year) (male%) (year) (months) Cryo RF Cryo RF... Kuck et al Germany RCT G or 2G, 3.5-mm irrigated s, 50 C, W, 30 s 23 or 28 mm tip, CF 2 5 times Luik et al Germany RCT NA 12 1G or 2G, 3.5-mm irrigated tip s, 2 times NA 23 or 28 mm Hunter et al UK RCT G, 23 or 28 mm 3.5-mm irrigated tip 300 s, 40 C, 30W 2 times Pérez-Castellano 2014 Spain RCT NA 12 1G, 23 or 28 mm 3.5-mm irrigated tip 300 s, 2 times 35W, 45 C et al. 19 Kardos et al Hungary OS, R NA 24 2G, 28 mm 3.5-mm irrigated tip, CF Providencia et al UK OS, P G, 23 or 28 mm 3.5 or 4 mm, irrigated tip 240 s, 1 time 25 35W, 48 C, 10 50g 240 s, 2 times 25 30W Khoueiry et al France OS, P G or 2G, 28 mm Irrigated tip, CF 240 s, 2 times 25 30W, 48 C Schmidt et al Germany OS, P NA 12 1G, 23 or 28 mm NA s NA Straube et al Germany OS, P NA 12 1G or 2G, 28 mm 3.5 or 4 mm, irrigated NA NA tip Squara et al France OS, P/R G, 23 or 28 mm CF 240 s, 2 times 20 35W, 38.5 C, 10 50g Wasserlauf et al USA OS, P G or 2G, 23 or 28 mm 3.5-mm irrigated tip s, 2 times 20 35W, 42 C, 20 40s Julia et al Belgium OS, R NA 12 1G or 2G, 28 mm 3.5-mm irrigated tip 240 s, 2 times 25 35W Jourda et al France OS, P G Irrigated tip, CF 240 s, 2 times 25 35W, 48 C, 10 50g Knecht et al Switzerland OS, P G, 23 mm 3.5-mm irrigated tip 300, 4 times 25 35W, 50 C Mugnai et al Belgium OS, R G, 28 mm 3.5-mm irrigated tip 300 s, 2 times 25 35W, 48 C, 60s Kühne et al Switzerland OS, P G, 28 mm 3.5-mm irrigated tip 300 s, 2 times 25 35W, 50 C Cryo, cryoablation; RF, radiofrequency ablation; RCT, randomized controlled trail; OS, observational study; P, perspective; R, retrospective; 1G, first generation cryoballoon; 2G, second generation cryoballoon; CF, contact force-sensing catheter; NA, none available.

5 788 Y.-H. Chen et al. Table 2 The baseline demographic characteristics of included patients No. of studies Cryoablation RF P value... Age (years) Gender (male%) AF history (years) Body mass index (kg/m 2 ) Left atrial diameter (mm) Hypertension (%) Coronary artery disease (%) Diabetes mellitus (%) Chronic kidney disease (%) CHA 2 DS 2 -VASc Stroke or TIA (%) LVEF (%) HAS-BLED Hyperlipidermia (%) RF, radiofrequency ablation; AF, atrial fibrillation; TIA, transit ischemic attack; LVEF, left ventricular ejection fractions. Study Events, Events, % ID RR (95% CI) cryoablation RF Weight Kuck et al. (2016) 1.02 (0.95, 1.10) 294/ / Luik et al. (2015) 1.04 (0.89, 1.21) 106/ / Hunter et al. (2015) 1.43 (1.07, 1.90) 52/78 36/ P rez Castellano et al. (2014) 0.71 (0.43, 1.15) 12/25 17/ Kardos et al. (2016) 1.03 (0.84, 1.27) 32/40 45/ Providencia et al. (2016) 1.72 (1.44, 2.06) 190/ / Khoueiry et al. (2016) 0.97 (0.91, 1.03) 258/ / Schmidt et al. (2016) 0.99 (0.91, 1.08) 329/ / Straube et al. (2016) 1.17 (0.96, 1.43) 76/107 60/ Squara et al. (2015) 0.96 (0.86, 1.09) 130/ / Wasserlauf et al. (2015) 0.99 (0.79, 1.24) 61/101 61/ Juli et al. (2015) 1.08 (0.94, 1.24) 77/ / Jourda et al. (2014) 0.97 (0.86, 1.10) 64/75 66/ Knecht et al. (2014) 0.85 (0.62, 1.17) 34/71 40/ Mugnai et al. (2014) 1.10 (0.94, 1.30) 86/ / K ¹hne et al. (2010) 0.96 (0.79, 1.15) 22/25 23/ Overall (I squared = 72.5%, p = 0.000) 1.05 (0.98, 1.13) 1823/ / NOTE: Weights are from random effects analysis Favour RF Favour cryoablation Figure 2 Forest plot of AF/atrial tachycardia-free survival for cryoablation vs. RF ablation. P ¼ 0.159), with low heterogeneity (I 2 ¼26.9, P ¼ 0.159) (Figure 5). Among the summarized adverse events, death from any cause was only occurred in 4 patients treated with cryoablation (0.1%). PNP occurred more frequently in cryoablation group (3.3%), which accounted for 41% of overall adverse events caused by cryoenergy application. Notwithstanding, most of them restored immediately while interrupting the free cycle or before discharge, with very few remained during the follow-up. Conversely, patients with RF ablation

6 Cryoablation vs. radiofrequency ablation for paroxysmal AF 789 Table 3 Univariable meta-regression analysis Covariates AF/atrial tachycardia-free survival Procedure time Fluoroscopy time Coefficient 95% CI P value Coefficient 95% CI P value Coefficient 95% CI P value... Age (years) ; ; ; Male (%) ; ; ; Sample size ; ; ; Follow-up (months) ; ; ; Study design ; ; ; CF-sensing ; ; ; catheter Cryoballoon generation ; ; ; Impact of study-level covariates on the relationship between the source of energy (cryoablation vs. RF ablation) and the efficacy and safety endpoints (AF/atrial tachycardia-free survival, procedure time and fluoroscopy time). 0 Standard Error of Relative Risk had a low risk of PNP (0.1%) compared with cryoablation. With respect to radiofrequency ablation, there was a higher prevalence of pericardial complications (2.1%), vascular complication (2.2%) and bleeding (2.4%) during the ablation, accounting for 27%, 27% and 30% of overall adverse events, respectively. A similar incidence of stroke or TIA was reported in both groups, with comparable ratio of corresponding total adverse events (4% for each). In addition to the above-mentioned complications, patients were also subjected to other rare adverse events (i.e. shock, coronary gas emboli, esophageal ulcer, contrast media reaction or psychotic symptoms). Overall information was detailed in Table 5. Discussion Relative Risk Figure 3 Funnel plot for assessment of publication bias for the AF/atrial tachycardia-free survival. In this systemic review and meta-analysis involving patients with paroxysmal AF, primary clinical endpoint of long-term AF/atrial tachycardia-free survival was found to be comparable between cryoablation and RF ablation. Consistently, subgroup analysis with respect to study design, sample size, use of CF-sensing catheter or generation of cryoballoon, as well as follow-up duration also showed similar results. Not surprisingly, we found a marked benefit of cryoablation in reducing procedure time when compared with RF ablation, whereas no evidence of significant difference in fluoroscopy time was observed between the two approaches. With regard to procedurerelated adverse events, pooled results showed an comparable overall adverse events in cryoablation to RF ablation. Cryoablation had emerged as a promising alternative approach to traditional point-to-point RF ablation, and was characterized by creating circumferential lesions with cryoballoon (23 or 28 mm) in a single shot. 28,29 In contrast to RF energy which resulted in cellular necrosis by tissue heating, irreversible injury to cells by cryothermal energy delivery were complex and attributed to the freeze and thaw cycle During the freezing process, ice crystal formation and osmotic gradient development led to acute cell death. Subsequent thawing process, the melting ice crystal caused cell swelling and disruption of membrane integrity by means of further osmotic insult. This histopathology feature may be responsible for the relatively demarcated lesions with minimal tissue architectural disruption in the setting of cryoablation, which contributed to lower risk of thrombus formation or perforation. 7,33 In addition to the advantage in cryolesions, cryoablation was particularly noted for the firmly adhesion to pulmonary veins and thus prevented from dislodgement. Meanwhile, increasingly utilization of cryoablation had attracted great attention into the underlying mechanism of cryothermal energy for the purpose of achieving desirable outcomes and avoiding complications. Recently, a well designed animal research by Takami et al. provided an elaboration on the biophysics of cryoablation. The main findings of this study were the correlation between the thermodynamics and distance of the cryoballoon from the tissue and the blood flow leak, along with the gap location of chronic reconduction. 34 Therefore, it offered operators a more deeply understanding of the efficacy and safety profiles of cryoablation and guided the application in clinical practice. The aforementioned characteristics of cryoablation theoretically fitted the expectation of operators and were considered as a potent

7 790 Y.-H. Chen et al. Table 4 Subgroup analysis of pooled estimates of AF/atrial tachycardia-free survival Subgroup NO. of studies Cryoablation... RF... RR (95% CI) P value Events Patients Events Patients... Study design RCT (0.91, 1.23) OS (0.96, 1.14) Sample size Larger (0.99, 1.17) Smaller (0.86, 1.14) CF-sensing catheter With (0.94, 1.03) Without (0.97, 1.22) Follow-up duration >12 months (0.94, 1.22) months (0.96, 1.10) Generation of cryoballoon 1G or 2G (0.97, 1.06) G only (0.90, 1.33) G only (0.90, 1.06) Total (0.98, 1.13) RF, radiofrequency ablation; RR, relative risk; CI, confidence interval; RCT, randomized controlled trail; OS, observational study; 1G, first generation cryoballoon; 2G, second generation cryoballoon. challenge to RF ablation for treating AF. A great deal of clinical studies had demonstrated the role of cryoablation in the treatment of either paroxysmal or persistent AF. 28,29,35 However, the pathogenesis of persistent AF was more complex and still remained unknown, and the ablation strategy was not only depended on the standard PVI but also the other adjunctive substrate modification (i.e. linear ablation, ablation of CFAE or ganglionic plexi) It may confuse the contribution of PVI and further influence the interpretation of effect for cryoablation in the treatment. On this account, our meta-analysis exclusively shone a spotlight on the efficacy and safety of cryoablation in paroxysmal AF compared with the conventional RF ablation. Previous observational studies indicated that long-term freedom from AF/atrial tachycardia ranged from 47.8% to 85.3% in cryoablation, which was comparable with RF ablation ,28,29,35 Recently, results of two large RCTs (the FreezeAF trail and the FIRE AND ICE trail), with regard to the comparison between the two ablation approaches, had been consecutively published. Both of studies found that the cryoablation was non-inferior to RF ablation in terms of the primary efficacy endpoint during a 12-month follow-up. 15,16 In accordance with this, pooled results in present meta-analysis showed 65.4% AF/atrial tachycardia-free survival of cryoablation, which was also similar to that of RF ablation (59.9%) in a median follow-up period of 12 months. With the advent of CF-sensing catheter and release of the second generation cryoballoon, nonetheless, it was still unknown whether either of them displayed a marked advantage in improving outcomes because of lacking sufficient clinical evidence. 7,14 Therefore, we conducted a subgroup analysis according to the use of CF-sensing catheter and the generation of cryoballoon. Neither of stratification significantly influenced the role of cryoablation in comparison with RF ablation. Furthermore, pooled results were also consistent when stratified by study design, sample size or follow-up duration and thus confirmed the robustness of our finding. In terms of safety during the ablation, cryoablation dramatically shortened the procedure time due to single-shot technology, which was in line with the majority of previous studies. 18,21 Intriguingly, there was no significant difference in fluoroscopy time between cryoablation and RF ablation. In cryoablation, angiography was routinely performed and a high image resolution was necessary in order to verify the occlusion of pulmonary vein. However, there was a series of studies providing contradictory results about the effect of cryoablation in fluoroscopy time. 15,22,27 Especially the study conducted by Wasserlauf et al., 22 which suggested a preferable effect of cryoablation with 27-min reduction in fluoroscopy time and was attributed to the potential impact of the learning curve and experience. Thus, it was of great importance to assess the actual effect of cryoablation in fluoroscopy time in the future. Additionally, procedure-related adverse events were of paramount importance during the catheter ablation. 39 Our meta-analysis demonstrated that the overall rate of adverse events was similar between the two approaches. To the best of our knowledge, PNP was the most common complication in patients treated with cryoablation, especially the use of second generation cryoballoon. 40,41 Unsurprisingly, the incidence of PNP was 3.3% and accounted for 41% of overall adverse events in cryoablation, in despite of excellent prognosis. Of note, another well-designed RCT (the STOP AF trail) reported a 11.2% rate of PNP occurred in cryoablation group which was significantly higher than that in our analysis. 42 The steady improvement in skills and cumulative experience may contribute to the reduction of PNP. In contrast, pericardial complication (2.1%), vascular complication (2.2%) and bleeding (2.4%) were frequently

8 Cryoablation vs. radiofrequency ablation for paroxysmal AF 791 A Study ID WMD (95% CI) % Weight Kuck et al. (2016) ( 23.31, 9.69) 6.93 Luik et al. (2015) ( 23.82, 2.18) 6.82 Hunter et al. (2015) ( 61.37, 26.63) 6.55 P rez Castellano et al. (2014) (9.73, 74.27) 5.66 Kardos et al. (2016) ( 59.67, 32.33) 6.72 Providencia et al. (2016) ( 22.28, 9.72) 6.94 Khoueiry et al. (2016) (12.69, 23.31) 6.96 Straube et al. (2016) ( 76.97, 59.03) 6.88 Squara et al. (2015) ( 21.07, 4.73) 6.90 Wasserlauf et al. (2015) ( , 73.27) 6.55 Juli et al. (2015) ( 86.55, 58.25) 6.70 Jourda et al. (2014) (10.62, 36.98) 6.74 Knecht et al. (2014) 1.00 ( 15.66, 13.66) 6.68 Mugnai et al. (2014) ( 91.42, 68.58) 6.80 K ¹hne et al. (2010) ( 54.93, 7.07) 6.20 Overall (I squared = 97.7%, p = 0.000) ( 45.24, 10.08) NOTE: Weights are from random effects analysis B Study Favour cryoablation Favour RF % ID WMD (95% CI) Weight Kuck et al. (2016) 5.10 (2.81, 7.39) 8.23 Luik et al. (2015) 0.50 ( 2.32, 3.32) 7.93 Hunter et al. (2015) 7.00 (1.96, 12.04) 6.41 P rez Castellano et al. (2014) 0.00 ( 8.87, 8.87) 4.04 Kardos et al. (2016) 2.00 ( 7.42, 3.42) 6.13 Providencia et al. (2016) 2.00 (0.52, 3.48) 8.60 Khoueiry et al. (2016) 2.30 (0.85, 3.75) 8.61 Straube et al (2016) 0.00 ( 2.12, 2.12) 8.32 Squara et al. (2015) 2.20 ( 4.18, 0.22) 8.39 Wasserlauf et al. (2015) ( 34.34, 19.66) 4.88 Juli et al. (2015) 8.60 ( 12.85, 4.35) 6.96 Jourda et al. (2014) 3.80 (0.85, 6.75) 7.85 Knecht et al. (2014) 8.00 ( 1.87, 17.87) 3.57 Mugnai et al. (2014) 5.00 ( 8.33, 1.67) 7.61 K ¹hne et al. (2010) (1.95, 28.05) 2.47 Overall (I squared = 89.4%, p = 0.000) 0.37 ( 2.78, 2.04) NOTE: Weights are from random effects analysis Favour cryoablation Favour RF Figure 4 Forest plot of fluoroscopy time (A) and procedure time (B) for cryoablation vs. RF ablation.

9 792 Y.-H. Chen et al. Study Events, Events, % ID RR (95% CI) cryoablation RF Weight Kuck et al. (2016) 0.85 (0.54, 1.33) 32/374 38/ Luik et al. (2015) 2.42 (1.09, 5.37) 19/156 8/ Hunter et al. (2015) 0.99 (0.26, 3.81) 4/78 4/ P rez Castellano et al. (2014) 1.00 (0.07, 15.12) 1/25 1/ Kardos et al. (2016) 4.35 (0.47, 40.33) 3/40 1/ Providencia et al. (2016) 1.19 (0.68, 2.08) 23/393 23/ Khoueiry et al. (2016) 0.97 (0.55, 1.71) 20/311 25/ Schmidt et al. (2016) 0.76 (0.55, 1.06) 41/ / Straube et al. (2016) 0.87 (0.48, 1.58) 19/189 19/ Squara et al. (2015) 1.03 (0.50, 2.14) 13/178 14/ Wasserlauf et al. (2015) 0.66 (0.19, 2.27) 4/101 6/ Jourda et al. (2014) 7.00 (1.65, 29.74) 14/75 2/ Knecht et al. (2014) 1.00 (0.21, 4.79) 3/71 3/ Mugnai et al. (2014) 1.34 (0.85, 2.12) 26/136 37/ K ¹hne et al. (2010) 1.00 (0.07, 15.12) 1/25 1/ Overall (I squared = 26.9%, p = 0.159) 1.08 (0.86, 1.35) 223/ / NOTE: Weights are from random effects analysis Favour cryoablation Favour RF Figure 5 Forest plot of procedure-related adverse events for cryoablation vs. RF ablation. Table 5 Overall procedure-related adverse events in the meta-analysis Adverse events Cryoablation... RF... Events Incidence (%) Ratio (%) Events Incidence (%) Ratio (%)... Death Pericardial PNP Stroke/TIA Vascular Bleeding Other Total RF, radiofrequency ablation; PNP, phrenic nerve palsy; TIA, transit ischemic attack. occurred in RF ablation and composed almost all adverse events (accounted for 27%, 27% and 30%, respectively). However, cryoablation was associated with a substantially lower rate of pericardial complication (0.8%) when compared with RF ablation, which may be attributed to the aforementioned characteristics of cryoenergy in preserving endocardial integrity during the application. 33 Moreover, other adverse events caused by cryoablation including esophageal injury, bronchial complication or even coronary gas emboli. Therefore, it can never be too cautious to perform catheter ablation with cryoballoon, notwithstanding the simplified and shortened pulmonary vein procedure. Study limitations First, our meta-analysis based on study level data rather than individual patient-level data. It was thus unable to explore the influence of other variables (i.e. history of paroxysmal AF or comorbidities) on the AF/atrial tachycardial-free survival, which may provide clinicians a novel perspective. Second, there was an evidence of high

10 Cryoablation vs. radiofrequency ablation for paroxysmal AF 793 heterogeneity for the pooled estimates of AF/atrial tachycardial-free survival. Sensitivity analysis revealed that the study conducted by Providencia et al. was the main source of heterogeneity with no impact on pooled RR and P value. 10 It was designed as a multicentre prospective cohort, in which enrolled patients from six hospitals. Nevertheless, significantly different annual caseloads and inappropriate assignment of patients may compromise the results and were the potential reason for the clinical heterogeneity in this study. Although the high heterogeneity may weaken the conclusions, robust sensitivity analysis with consistent results could overcome the deficiency. Third, it should be cautious in interpreting and extending our results due to the variation in ablation parameters, catheter type and even the geographic region. Finally, the majority of studies included in our analysis were observational, with a higher risk of selection bias. Therefore, there was a pressing need of well designed RCTs in the future. Conclusions In patients with paroxysmal AF, cryoablation showed a comparable efficacy in maintaining sinus rhythm vs. RF ablation at a median 12 months of follow-up. In addition, cryoablation significantly shortened the procedure time, nonetheless, with negligible impact on the fluoroscopy time. Moreover, there was no significant difference in adverse events between the two approaches. With the increasingly use of cryoballoon, more RCTs were needed to further assess the role of cryoablation and guided clinical practice for achieving optimal outcomes. Acknowledgements We thank Professor Y.-G.L. and H.L. for assisting with the guiding and revising of the article. We also thank all the study participants. Conflict of interest: none declared. References 1. Camm AJ, Lip GY, De Caterina R, Savelieva I, Atar D, Hohnloser SH et al focused update of the ESC Guidelines for the management of atrial fibrillation: an update of the 2010 ESC Guidelines for the management of atrial fibrillation. Europace 2012;14: January CT, Wann LS, Alpert JS, Calkins H, Cigarroa JE, Cleveland JC et al AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. J Am Coll Cardiol 2014;64:e Calkins H, Kuck KH, Cappato R, Brugada J, Camm AJ, Chen SA et al HRS/ EHRA/ECAS Expert Consensus Statement on Catheter and Surgical Ablation of Atrial Fibrillation: recommendations for patient selection, procedural techniques, patient management and follow-up, definitions, endpoints, and research trial design. Europace 2012;14: Ouyang F, Tilz R, Chun J, Schmidt B, Wissner E, Zerm T et al. Long-term results of catheter ablation in paroxysmal atrial fibrillation: lessons from a 5-year followup. Circulation 2010;122: Kojodjojo P and Wyn Davies D. How to perform antral pulmonary venous isolation using the cryoballoon. Heart Rhythm 2011;8: Baman TS, Jongnarangsin K, Chugh A, Suwanagool A, Guiot A, Madenci A et al. Prevalence and predictors of complications of radiofrequency catheter ablation for atrial fibrillation. J Cardiovasc Electrophysiol 2011;22: Wissner E, Deiss S and Kuck KH. Balloon-based pulmonary vein isolation for the treatment of paroxysmal atrial fibrillation (cryoballoon and laserballoon). Future Cardiol 2015;11: Wojcik M, Berkowitsch A, Greis H, Zaltsberg S, Hamm CW, Pitschner HF et al. Learning curve in cryoballoon ablation of atrial fibrillation: eight-year experience. Circ J 2014;78: Mikhaylov EN, Lebedev DS, Pokushalov EA, Davtyan KV, Ivanitskii EA, Nechepurenko AA et al. Outcomes of cryoballoon ablation in high- and lowvolume atrial fibrillation ablation centres: a Russian Pilot Survey. Biomed Res Int 2015;2015: Providencia R, Defaye P, Lambiase PD, Pavin D, Cebron JP, Halimi F et al. Results from a multicentre comparison of cryoballoon vs. radiofrequency ablation for paroxysmal atrial fibrillation: is cryoablation more reproducible? Europace 2016 [Epub ahead of print]. 11. Kardos A, Kis Z, Som Z, Nagy Z and Foldesi C. Two-year follow-up after contact force sensing radiofrequency catheter and second-generation cryoballoon ablation for paroxysmal atrial fibrillation: a comparative single centre study. Biomed Res Int 2016;2016: Khoueiry Z, Albenque JP, Providencia R, Combes S, Combes N, Jourda F et al. Outcomes after cryoablation vs. radiofrequency in patients with paroxysmal atrial fibrillation: impact of pulmonary veins anatomy. Europace 2016;18: Reddy VY, Dukkipati SR, Neuzil P, Natale A, Albenque JP, Kautzner J et al. Randomized, controlled trial of the safety and effectiveness of a contact forcesensing irrigated catheter for ablation of paroxysmal atrial fibrillation: results of the tacticath contact force ablation catheter study for atrial fibrillation (TOCCASTAR) study. Circulation 2015;132: Su W, Kowal R, Kowalski M, Metzner A, Svinarich JT, Wheelan K et al. Best practice guide for cryoballoon ablation in atrial fibrillation: the compilation experience of more than 3000 procedures. Heart Rhythm 2015;12: Kuck KH, Brugada J, Fürnkranz A, Metzner A, Ouyang F, Chun KR et al. Cryoballoon or radiofrequency ablation for paroxysmal atrial fibrillation. N Engl J Med 2016;374: Luik A, Radzewitz A, Kieser M, Walter M, Bramlage P, Hormann P et al. Cryoballoon versus open irrigated radiofrequency ablation in patients with paroxysmal atrial fibrillation: the prospective, randomized, controlled, noninferiority FreezeAF study. Circulation 2015;132: Squara F, Zhao A, Marijon E, Latcu DG, Providencia R, Di Giovanni G et al. Comparison between radiofrequency with contact force-sensing and secondgeneration cryoballoon for paroxysmal atrial fibrillation catheter ablation: a multicentre European evaluation. Europace 2015;17: Hunter RJ, Baker V, Finlay MC, Duncan ER, Lovell MJ, Tayebjee MH et al. Pointby-point radiofrequency ablation versus the cryoballoon or a novel combined approach: a randomized trial comparing 3 methods of pulmonary vein isolation for paroxysmal atrial fibrillation (the Cryo versus RF trial). J Cardiovasc Electrophysiol 2015;26: Perez-Castellano N, Fernandez-Cavazos R, Moreno J, Canadas V, Conde A, Gonzalez-Ferrer JJ et al. The COR trial: a randomized study with continuous rhythm monitoring to compare the efficacy of cryoenergy and radiofrequency for pulmonary vein isolation. Heart Rhythm 2014;11: Schmidt M, Dorwarth U, Andresen D, Brachmann J, Kuck K, Kuniss M et al. German ablation registry: cryoballoon vs. radiofrequency ablation in paroxysmal atrial fibrillation one-year outcome data. Heart Rhythm 2016;13: Straube F, Dorwarth U, Ammar-Busch S, Peter T, Noelker G, Massa T et al. First-line catheter ablation of paroxysmal atrial fibrillation: outcome of radiofrequency vs. cryoballoon pulmonary vein isolation. Europace 2016;18: Wasserlauf J, Pelchovitz DJ, Rhyner J, Verma N, Bohn M, Li Z et al. Cryoballoon versus radiofrequency catheter ablation for paroxysmal atrial fibrillation. Pacing Clin Electrophysiol 2015;38: Julia J, Chierchia GB, de Asmundis C, Mugnai G, Sieira J, Ciconte G et al. Regular atrial tachycardias following pulmonary vein isolation for paroxysmal atrial fibrillation: a retrospective comparison between the cryoballoon and conventional focal tip radiofrequency techniques. J Interv Card Electrophysiol 2015;42: Jourda F, Providencia R, Marijon E, Bouzeman A, Hireche H, Khoueiry Z et al. Contact-force guided radiofrequency vs. second-generation balloon cryotherapy for pulmonary vein isolation in patients with paroxysmal atrial fibrillation-a prospective evaluation. Europace 2015;17: Knecht S, Sticherling C, von Felten S, Conen D, Schaer B, Ammann P et al. Longterm comparison of cryoballoon and radiofrequency ablation of paroxysmal atrial fibrillation: a propensity score matched analysis. Int J Cardiol 2014;176: Mugnai G, Chierchia GB, de Asmundis C, Sieira-Moret J, Conte G, Capulzini L et al. Comparison of pulmonary vein isolation using cryoballoon versus conventional radiofrequency for paroxysmal atrial fibrillation. Am J Cardiol 2014;113: Kuhne M, Suter Y, Altmann D, Ammann P, Schaer B, Osswald S et al. Cryoballoon versus radiofrequency catheter ablation of paroxysmal atrial fibrillation: biomarkers of myocardial injury, recurrence rates, and pulmonary vein reconnection patterns. Heart Rhythm 2010;7: Van Belle Y, Janse P, Rivero-Ayerza MJ, Thornton AS, Jessurun ER, Theuns D et al. Pulmonary vein isolation using an occluding cryoballoon for circumferential ablation: feasibility, complications, and short-term outcome. Eur Heart J 2007;28:

11 794 Y.-H. Chen et al. 29. Neumann T, Vogt J, Schumacher B, Dorszewski A, Kuniss M, Neuser H et al. Circumferential pulmonary vein isolation with the cryoballoon technique results from a prospective 3-center study. J Am Coll Cardiol 2008;52: Mazur P. Kinetics of water loss from cells at subzero temperatures and the likelihood of intracellular freezing. J Gen Physiol 1963;47: Lustgarten DL, Keane D and Ruskin J. Cryothermal ablation: mechanism of tissue injury and current experience in the treatment of tachyarrhythmias. Prog Cardiovasc Dis 1999;41: Kaszala K and Ellenbogen KA. Biophysics of the second-generation cryoballoon: cryobiology of the big freeze. Circ Arrhythm Electrophysiol 2015;8: Khairy P, Chauvet P, Lehmann J, Lambert J, Macle L, Tanguay JF et al. Lower incidence of thrombus formation with cryoenergy versus radiofrequency catheter ablation. Circulation 2003;107: Takami M, Misiri J, Lehmann HI, Parker KD, Johnson SB, Sarmiento RI et al. Spatial and time-course thermodynamics during pulmonary vein isolation using the second-generation cryoballoon in a canine in vivo model. Circ Arrhythm Electrophysiol 2015;8: Vogt J, Heintze J, Gutleben KJ, Muntean B, Horstkotte D, Nolker G. Long-term outcomes after cryoballoon pulmonary vein isolation: results from a prospective study in 605 patients. J Am Coll Cardiol 2013;61: Ernst S, Ouyang F, Lober F, Antz M and Kuck KH. Catheter-induced linear lesions in the left atrium in patients with atrial fibrillation: an electroanatomic study. J Am Coll Cardiol 2003;42: Nademanee K, McKenzie J, Kosar E, Schwab M, Sunsaneewitayakul B, Vasavakul T et al. A new approach for catheter ablation of atrial fibrillation: mapping of the electrophysiologic substrate. J Am Coll Cardiol 2004;43: Scherlag BJ, Nakagawa H, Jackman WM, Yamanashi WS, Patterson E, Po S et al. Electrical stimulation to identify neural elements on the heart: their role in atrial fibrillation. J Interv Card Electrophysiol. 2005;13: Mugnai G, Irfan G, de Asmundis C, Ciconte G, Saitoh Y, Hunuk B et al. Complications in the setting of percutaneous atrial fibrillation ablation using radiofrequency and cryoballoon techniques: a single-center study in a large cohort of patients. Int J Cardiol 2015;196: Mugnai G, de Asmundis C, Ciconte G, Irfan G, Saitoh Y, Velagic V et al. Incidence and characteristics of complications in the setting of secondgeneration cryoballoon ablation: a large single-center study of 500 consecutive patients. Heart Rhythm 2015;12: Furnkranz A, Bordignon S, Schmidt B, Perrotta L, Dugo D, De Lazzari M et al. Incidence and characteristics of phrenic nerve palsy following pulmonary vein isolation with the second-generation as compared with the first-generation cryoballoon in 360 consecutive patients. Europace 2015;17: Packer DL, Kowal RC, Wheelan KR, Irwin JM, Champagne J, Guerra PG et al. Cryoballoon ablation of pulmonary veins for paroxysmal atrial fibrillation: first results of the North American Arctic Front (STOP AF) pivotal trial. J Am Coll Cardiol 2013;61: