There are several main findings from our study. First, our study revealed that 2G-CB ablation had similar efficacy with CF-RF ablation for paroxysmal AF in terms of acute PVI, freedom from ATAs, and freedom from AF. Second, the 2G-CB ablation procedure can be performed faster than CF-RF ablation, even though the fluoroscopy time between both groups was similar. Third, 2G-CB ablation was correlated with a higher phrenic nerve paralysis rate. Moreover, all phrenic nerve paralysis complications happened in the 2G-CB group.
Today's paradigm is that electrical isolation of the pulmonary veins from the left atrium is the fundamental of most catheter-based ablation strategies in paroxysmal AF. However, there are no specific recommendations from the recent guidelines regarding the choice of CBA or RFA [9, 51, 52]. Until now, the largest RCT comparing CBA and RFA in paroxysmal AF is the FIRE AND ICE trial. The study revealed that CBA was not inferior to RFA regarding the efficacy. The overall safety of both procedures was not significantly different. In the FIRE AND ICE trial, the CBA procedures were conducted using 1G-CB or 2G-CB catheters. Moreover, the data of CF-RF catheter were not reported in that trial . The FreezeAF study also revealed the noninferiority of CBA than RFA for rhythm control in paroxysmal AF patients . A meta-analysis of RCTs from Murray et al. comparing CBA using 1G-CB or 2G-CB catheters and RFA demonstrated that CBA and RFA shared equal efficacy. However, that meta-analysis did not provide information about the use of CF-RF catheters . A meta-analysis from Jiang et al. revealed that 2G-CB ablation effectively decreased the recurrence rate of ATAs compared to RFA in paroxysmal AF patients specifically .
Buist et al. conducted RCT to compare 2G-CB ablation and CF-RF ablation in AF patients. However, that study included both paroxysmal AF and persistent AF. That study demonstrated that 2G-CB ablation provided better ATAs-free survival and lower repeat ablation than CF-RF ablation . The CIRCA-DOSE study revealed that 2G-CB ablation and CF-RF ablation resulted in similar efficacy for paroxysmal AF during a one-year follow-up duration . However, the study was compared longer and shorter application time of cryoballoon (CB) ablation to CF-RF ablation, which causes the CG group to be non-uniform. A meta-analysis from Ravi et al., including RCT and cohort studies comparing 2G-CB ablation and CF-RF ablation, revealed that the efficacy between both groups was similar . Another meta-analysis from Wang et al. that included RCTs showed that AF recurrence rate between 2G-CB ablation and CF-RF ablation were comparable . However, the meta-analysis study from Ravi et al. and Wang et al. involved both paroxysmal AF and persistent AF patients [56, 57]. Compared to the prior meta-analysis, our study specifically compared 2G-CB ablation and CF-RF ablation in patients with paroxysmal AF. Our study also revealed a similar success rate of acute PVI between groups. This result supported the previous study by Wang et al .
Our study demonstrated that 2G-CB ablation in paroxysmal AF could be completed faster than CF-RF ablation. Our result was consistent and supported the previous meta-analysis study from Ravi et al.  and Wang et al . The 2G-CB ablation can be conducted faster because of its “single-shot” characteristic used throughout the PVI. On the other hand, CF-RF ablation needs a longer procedure time because of its “point-by-point” approach . Previous meta-analysis demonstrated that fluoroscopy time was longer in 2G-CB ablation than CF-RF ablation . However, in our study, both groups revealed no significantly different fluoroscopy time. We found significant heterogeneity while conducting data analysis of procedure time and fluoroscopy time. That was because of the diverse habit and experience of fluoroscopy utilization among different heart rhythm centers. The increased experience of the operator to perform AF ablation could reduce the fluoroscopy time . The high power and short-duration (HPSD) radiofrequency ablation is now being conducted to reduce the overall procedure time in CF-RF ablation. A study from Baher et al. revealed that compared to the conventional method (35 W power for 10 to 30 seconds), the HPSD approach (50 W for 5 seconds) took a shorter procedure time (149 ± 65 minutes vs. 251 ± 101 minutes; p < 0.001) . Until now, no study specifically compared the 2G-CB ablation and HPSD CF-RF ablation in paroxysmal AF patients. Moreover, almost all CF-RF ablation procedures in this meta-analysis were conducted using the conventional method (25 to 35 W power for at least 20 seconds) [40–44, 46–50].
From the safety aspect, our study revealed that 2G-CB ablation and CF-RF ablation had not significantly different rates of all-procedural complication, pericardial effusion, and vascular complications. Our results supported the findings of prior studies. However, those meta-analyses did not provide the data about pericardial effusion and vascular complications [56–57]. Our result revealed that the incidence of pericardial effusion was not significantly different in both groups. However, in a prior meta-analysis from Jiang et al., CF-RF ablation had a higher rate of pericardial tamponade than RFA . The possible explanation was: (1). The meta-analysis from Jiang et al. included RFA using the non-CF-RF catheter and CF-RF catheter ; (2) Our meta-analysis only included CF-RF ablation; (3) In our meta-analysis, almost all studies used the low-power and/or long-duration radiofrequency strategy in the CF-RF group [41, 44, 46, 50] and (4) CF-RF catheter provides efficient transfer of heat energy to the ablation target . The risk of phrenic nerve paralysis in our meta-analysis was higher in the 2G-CB group than in the CF-RF group. Our result was similar and supported the findings of the prior meta-analysis studies [53, 56].
To the best of our knowledge, our study is the first systematic review and meta-analysis study comparing 2G-CB ablation and RF-CF ablation for paroxysmal AF patients specifically. We also did not find any publication bias in this study. The studies involved in this systematic review and meta-analysis were mainly cohort studies [41–47, 49, 50]. However, we only involved the high-quality study in this meta-analysis.
We recognized several limitations in this meta-analysis. First, the data about the specific comorbidities were not always completely available in most studies [39–44, 46–50]. Second, freedom of ATAs definition among the included studies was varied [39–50]. Third, even though almost all included studies used 12 lead ECG and Holter-monitor as the arrhythmia detection methods [39–49], two studies used additional methods such as external loop recorder and auto-triggered event monitor [47, 50]. The last, the differences in blanking period or follow-up durations and the use of AADs during those periods. Those limitations could be the essential confounders that would affect the final results.