Thoracoscopic surgical ablation or catheter ablation for patients with atrial fibrillation? A systematic review and meta-analysis of randomized controlled trials

Abstract

Open in new tabDownload slide

OBJECTIVES

There is an urgent need to understand the difference in the influence of thoracoscopic surgical ablation (TSA) and catheter ablation (CA) on clinical outcomes in patients with atrial fibrillation (AF). This meta-analysis of randomized controlled trials aimed to examine the efficacy and safety of TSA versus CA in patients with AF.

METHODS

Databases including EMBASE, Clinical Trials, PubMed and Cochrane Central Registered Control System were screened for the retrieval of articles. A direct meta-analysis of TSA versus CA was conducted. The I² test analysis was performed to evaluate heterogeneity. The Begg–Mazumdar test and the Harbord–Egger test were used to detect publication bias. The primary efficacy outcome was freedom from atrial tachyarrhythmia, while the primary safety outcome was severe adverse event (SAE) occurrence.

RESULTS

Of the 860 identified articles, 6, comprising 466 participants, were finally included. The rate of freedom from AT was higher in the TSA group (75%) than in the CA group (57.1%) (odds ratio 0.41; 95% confidence interval 0.2–0.85; P = 0.02; I² = 57%). A larger number of SAEs were observed in the TSA group than in the CA group (odds ratio 0.16; 95% confidence interval 0.006–0.46; P = 0.0006; I² = 44%). The result of the subgroup analysis of 3 studies that enrolled AF patients without a history of ablation showed that the incidence of AT was comparable in both arms. The ablation procedure and hospitalization durations were longer in the TSA arm.

CONCLUSIONS

In our study, TSA was associated with better efficacy but a higher rate of SAEs compared to CA. In addition, TSA did not show better efficacy results as the first invasive procedure in the sub-analysis of patients with paroxysmal AF or early persistent AF. Therefore, doctors should recommend either TSA or CA to patients with AF after due consideration of the aforementioned findings.

INTRODUCTION

In the past few decades, great progress has been made in the field of non-drug therapy for atrial fibrillation (AF). Several treatment strategies for patients with AF have been developed, including surgical approaches and the percutaneous interventional catheter approach [1]. Catheter ablation (CA) is considered to be effective and safe in restoring and maintaining the sinus rhythm in patients with AF; its use is recommended by clinical guidelines and it is currently widely used in clinical treatment [2, 3]. In terms of the surgical approach, various types of techniques are available [4–6], most of which are used in patients with AF with a secondary heart problem requiring cardiac surgery. Thoracoscopic surgical ablation (TSA) was believed to be a promising treatment for AF, owing to its minimally invasive nature [7].

During CA, pulmonary vein isolation (PVI) through the ablation of the endocardial wall of the left atrium and additional lines could be achieved with a lower rate of trauma [7], while during TSA, PVI could be achieved using epicardial stand-alone ablation with bipolar ablation clamp use. Along with PVI, the left atrial appendage could be removed, the Marshall ligament could be disconnected, epicardial ganglionic plexi could be ablated and a posterior box lesion could be created with the assistance of thoracoscopy [8]. Both methods have their own advantages and disadvantages.

As 2 promising treatment methods of AF, TSA and CA are usually performed by 2 different groups of doctors: 1 is performed by cardiac surgeons and the other is performed by cardiologists. Meanwhile, there is no conclusion as to which method is safer and more effective for AF. Although an increasing number of studies have been focusing on the comparison of CA and TSA, not all of them have provided similar results. A randomized clinical trial conducted by   Boersma et al. [9] showed that CA was inferior to TSA in the achievement of freedom from left atrial arrhythmia. Meanwhile, 2 recently published reports showed that TSA was not superior to CA in terms of efficacy and showed a greater rate of complications [10, 11].

Previously, 2 meta-analysis reviews focused on CA and TSA [6, 12]. However, in those reviews, most of the studies involved were of a retrospective nature, and several recently published randomized controlled trials (RCTs) were not included [10, 13, 14]. Accordingly, we sought to conduct a meta-analysis of RCTs comparing CA to TSA, to examine their efficacy and safety in patients with AF.

METHODS

Search strategy and study selection

Databases including EMBASE, Clinical Trials, PubMed and Cochrane Central Registered Control System were screened for the retrieval of articles till 30 December 2019. The following terms were used as MeSH or keywords: atrial fibrillation, paroxysmal atrial fibrillation, persistent atrial fibrillation, catheter ablation, endocardial ablation, cryoballoon ablation, radiofrequency ablation, thoracoscopy, thoracoscopic ablation and epicardial ablation. The detailed search strategy for one of the electronic database searches (PubMed) is shown in Supplementary Material, Table S1. This review and meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and consulted guidelines of meta-analysis for the European Journal of Cardiothoracic Surgery and the Interactive Cardiovascular and Thoracic Surgery [15]. After retrieving relevant literature, all titles were reviewed; then, the abstracts of selected citations were reviewed to identify the full text of the retrieved articles based on the inclusion and exclusion criteria. Finally, the qualification of the full text of potentially related articles was evaluated. The decision whether to include each study was made independently by 2 investigators.

Eligibility criteria

The following criteria were applied for study inclusion: published in English; contain original data regarding CA versus TSA in patients with AF; RCT design. The exclusion criteria were: conference abstracts; original data regarding hybrid surgery; reviews; lack of data on the major outcomes of interest; fewer than 10 patients in each arm; follow-up duration shorter than 6 months. Only the most complete report was included in the case of duplicates.

Data extraction and critical appraisal

If the incorporated articles passed the quality assessment, data were extracted. The following data, including those on study design, number of patients in each study arm, baseline characteristics, outcome measures of interest, procedural details of CA and TSA, and follow-up duration, were extracted from each study. Two authors independently extracted all the necessary data using the data extraction form. Disagreements were resolved through discussions to achieve consensus. Quality assessment of the included studies was performed using the Cochrane Risk of Bias Tool for RCTs [16].

Outcomes

The primary efficacy outcome in our research was freedom from atrial tachyarrhythmia without antiarrhythmic drug use. The primary safety outcome was the occurrence of severe adverse events (SAEs), including death, acute coronary syndrome, myocardial infarction, stroke, transient ischaemic attack, nerve damage, major bleeding requiring surgery/blood transfusion, vascular damage or insufficiency requiring treatment, cardiac tamponade/perforation, significant/symptomatic pulmonary vein stenosis, infection, pericarditis, atrioesophageal fistula, diaphragmatic paresis/paralysis persistent air leak, pneumothorax, hydrothorax/pneumothorax, empyema, acute coronary syndrome, myocardial infarction, superficial wound and conversion to complete thoracotomy. Secondary outcomes included the ablation procedure and hospitalization durations.

Statistical analysis

Weighted mean differences (MDs) and 95% confidence intervals (CIs) were used to express continuous variables. Heterogeneity was assessed by I². If the I² value was ≤40%, a fixed-effect model was used, whereas if it was >40%, both the fixed-effect and random-effects models were used accordingly [17]; only the result of the random-effects model would be displayed. If the degree of heterogeneity was greater, sample-sensitivity analysis was conducted through the removal of 1 study each time. Values presented as median and interquartile range in the involved studies were converted to mean and standard deviation [18]. Analyses were performed using the Review Manager Version 5.3 (The Cochrane Collaboration, Copenhagen, Denmark). P-values <0.05 were considered statistically significant. To assess publication bias, the Begg–Mazumdar test [19] and the Harbord–Egger test [20] were used with Stata version 12.0 (Stata Corporation, College Station, TX, USA). P < 0.05 indicated the presence of publication bias.

RESULTS

Study selection

As shown in Fig. 1, 860 articles were identified through the database search. In total, 261 duplicates were removed and 23 articles remained after title and abstract screening. During the full-text screening, 5 studies were excluded due to their non-RCT design [11, 21–24], 9 were excluded due to the absence of data on results [25–33], 2 were excluded as they were duplicates with incomplete data [34, 35], and 1 was excluded as it was a conference abstract [36]. Six articles were finally included in the study [9, 10, 13, 14, 37, 38].

As shown in Table 1, all the studies involved were RCTs; 3 of them only enrolled patients with PAF [13, 14, 38], while the remaining 3 enrolled those with PAF and persistent AF (PersAF) [9, 10, 37]. The study by   Pokushalov et al. [37] only included patients with a history of failed CA, while that by Lucas et al. [9] and Sugihara et al. [13] enrolled those with/without a history of CA. The other 3 studies only included patients without a history of CA [11, 15, 38]. One study randomly assigned the enrolled patients to 3 groups: the pulmonary vein ablation catheter group, circular, irrigated mapping (nMARQ) CA group and TSA group [13]; the data from the pulmonary vein ablation catheter and nMARQ groups were combined during data extraction, as both of them used CA. No repeat ablation was performed in 5 studies [9, 10, 14, 37, 38]. In the study by Sugihara et al. [13], a repeat ablation was involved. For the detection of AF recurrence, electrocardiography and a Holter monitor test were used in 3 studies [9, 14, 38], and an implanted monitoring device was used in the other 3 [10, 13, 37]. In 2 studies, the recurrence of AF was defined as AF lasting >30 s [34, 38]; in 1 study, it was defined as AF% > 0.5% [37]; in another study, it was defined as detection of the presence of irregular R–R intervals lasting >2 min [13]. The primary efficacy end point was defined as freedom from any LA arrhythmia lasting >30 s in 2 studies [9, 10]. A longer follow-up duration of 24 months was observed in the TSA group in the study by  Adiyaman et al. [10], compared to the 12-month follow-up period observed in the other studies [9, 13, 14, 37, 38].

A total of 466 participants were included in our research: 254 participants were present in the CA arm and 212 in the TSA arm. The mean age of all the patients was 56.1 ± 9.3 years, 38.7% of the participants showed PAF, and the percentage of men was 68.2, as shown in Table 2. The duration of preoperative AF was 5.8 ± 4.5 years in the CA group, compared to 6.3 ± 5.7 in the TSA group.

The Cochrane’s risk of bias tool was used to evaluate the involved studies, and the results are presented in Fig. 2. In terms of allocation bias, 2 studies did not explicitly state the random approach with an unclear risk of bias [13, 38], and random sequence generation was detailed in 4 studies with a low risk of bias [9, 10, 14, 37]. In terms of the nature of the ablation procedures, it was not possible to blind patients to their treatment methods. The assumption was that the blinding measure would not affect outcomes and that the blinding of outcome assessments entailed a low risk in the involved studies. There was no drop out in most of the included studies [9, 14, 37, 38], with the exception of the studies by Ahmet et al. [10] and   Sugihara et al. [13], which showed drop-out rates of 5.8% and 6.9%, respectively. In 2 studies, all the predefined outcomes were reported [10, 14], while some prespecified outcomes were not found in the remaining 4 studies [9, 13, 37, 38]. Regarding other potential biases, 2 of the studies were considered as having a low risk [37, 38], 3 showed an unclear risk associated with the small simple size [10, 13, 14] and 1 showed an unclear risk in relation to its industry funding [9]. For the publication bias, Begg’s test gave a value of Pr > |z| = 0.806, and Egger’s test gave a value of p > |t| = 0.725, suggesting no significant publication bias for the included studies.

Primary outcomes

All 6 studies and 466 participants were included in the analysis of the primary efficacy outcome. The incidence of freedom from atrial tachyarrhythmia was higher in the TSA group (75%) than in the CA group (57.1%) [odds ratio (OR)  0.41; 95% CI 0.2–0.85; P = 0.02; I² = 57%] (Fig. 3A). To identify the reason for the higher heterogeneity, sensitivity analysis was conducted, with the results attributing it to the study by Ahmet et al. [10], as the I² value became 0% when this study was removed (Table 3). Subgroup analysis of 3 studies that enrolled AF patients without a history of ablation was also conducted [10, 14, 38]. The results showed that the incidence of freedom from atrial tachyarrhythmia was comparable in the TSA and CA arms (OR 0.69; 95% CI 0.14–3.38; P = 0.64; I² = 72%).

As shown in Fig. 3B, 5 studies and 328 participants were involved in the analysis of SAE occurrence [9, 10, 13, 14, 37]. A larger number of SAEs occurred in the TSA group than the CA group (OR = 0.16, 95% CI = 0.006–0.46; P = 0.0006; I² = 44%). The result of the subgroup analysis of 2 studies that enrolled AF patients without a history of ablation also showed that a larger number of SAEs occurred in the TSA group (OR = 0.2, 95% CI = 0.06–0.72; P = 0.01; I² = 0%) [10, 13, 14, 38]. There was 1 death in CA group and 1 death in TSA group.

Secondary outcomes

As shown in Fig. 4, 5 studies were involved in the analysis of procedure duration [9, 10, 13, 37, 38] and 4 were involved in the analysis of hospitalization duration [10, 13, 14, 37]. The results showed that the procedure duration and hospitalization in the TSA arm were longer than those in the CA arm (MD −74.44; 95% CI −142.06 to −6.82; P = 0.03; I² = 98%; MD −4.3; 95% CI −6.59 to −2.01; P = 0.002; I² = 96%, respectively). The results pertaining to the durations of ablation and hospitalization in all the involved studies showed the same trend favouring the CA group; the very high heterogeneity observed during the analyses may be attributed to the different strategies employed and the experiences of the operators in different centres. Individual studies did not influence the overall risk estimate of the durations of ablation and hospitalization.

The quality of evidence, as measured by the GRADE system, is shown in Table 4.

DISCUSSION

To the best of our knowledge, this is the first meta-analysis and review of RCTs to compare the efficacy and safety of TSA and CA. Our results showed that the incidence of freedom from atrial tachyarrhythmia was higher in the TSA group than in the CA group, while the occurrence rate of SAEs was higher in the TSA group during the follow-up. The ablation and hospitalization durations were longer in the TSA arm than in the CA arm. The subgroup analysis of studies that used ablation as the first invasive procedure showed that the TSA group had a similar success rate but a higher incidence of complications than the CA group.

The incidence of freedom from atrial tachyarrhythmia was 75% in the TSA group and 57.1% in the CA group in our study, indicating that TSA was significantly more efficacious than CA in reducing the recrudescence of AF in all the participants. During both invasive procedures, PVI was the cornerstone of ablation [39]. In addition to PVI, the posterior left atrial box lesion could be created, left atrial appendage could be excluded, ganglionic plexi could be ablated and the Marshall ligament could be resected during TSA, adding to the efficacy-related benefit of TSA in the elimination of other potential mechanisms that could induce atrial arrhythmia [9, 12, 40]. Meanwhile, it was very challenging to obtain contiguous, transmural, unrecovered, and permanent atrial lesions by the endocardial approach during CA [37, 41]. Current literature indicates that the presence of a larger number of interventions may yield a higher ablation success rate in AF [8, 42]. In addition, a greater number of intervention strategies could be facilitated during TSA performance.

Although TSA is distinct from complete open chest Cox-maze surgery and is considered minimally invasive, general anaesthesia, sequential lung deflation and transthoracic access were still required during TSA, which lead to process-related complications [9, 43]. In contrast, CA only required a small femoral puncture hole for the creation of access to the heart under local anaesthesia. As expected, the occurrence of SAEs was higher in the TSA group, accompanied by longer procedure and hospitalization durations, than in the CA group.

To identify the most efficient first invasive treatment for AF, subgroup analyses of 3 studies that enrolled patients without a history of previous ablation were conducted. All the involved studies included patients with PAF or early PersAF (continuous duration <3 months). The results showed that TSA as the first invasive measure was not superior to CA in the achievement of freedom from atrial arrhythmia in patients with PAF or early PersAF. Previous research has demonstrated that 94% of the focal sources in patients with PAF were located in the pulmonary vein [44]. Several studies subsequently demonstrated that, as the cornerstone, PVI could prevent the recurrence of AF in most patients with PAF or early PersAF [45]. In addition, for long-standing PersAF and cases with AF recurrence, further ablation strategies are required [46]. The fact that TSA did not yield better efficacy in the sub-analysis of patients with PAF or early PersAF could be due to the pulmonary vein origin being the main cause in such patients; PVI was sufficient for the prevention of AF occurrence in most of them. The ability of TSA to allow for the performance of a larger number of intervention strategies was not reflected in AF patients undergoing first ablation.

Our findings strongly suggest that doctors should exercise caution in the recommendation of either TSA or CA to patients with AF, as, compared to CA, TSA shows higher efficacy along with a greater rate of SAEs. Several factors need to be considered, including the AF type, AF duration and whether the invasive procedure is being used for the first time since TSA did not show better efficacy in the sub-analysis of patients with PAF or early PersAF as the first invasive procedure, with a higher rate of procedure-related complications. Therefore, future research must investigate if specific subgroups of AF patients would benefit to a greater degree from CA or TSA [47].

Although 2 meta-analyses focusing on CA and TSA have been conducted previously [6, 12], ours is the first meta-analysis to include only RCTs—the ‘gold standard’ for clinical trials [48]. In the aforementioned meta-analyses, most of the involved studies had a retrospective design. Therefore, the results of those non-RCT studies included may be non-homogeneous, leading to their unreliability. Most importantly, several recently published RCTs were not included in the published meta-analyses [10, 13, 14], affecting the reliability of their conclusions.

A high degree of heterogeneity (I² = 57%) was observed during the analysis of the primary efficacy outcome. In the sensitivity analysis, the high heterogeneity was shown to be caused by the study by Ahmet et al. [10], as the I² value became 0% after its removal. Several differences were observed between the studies conducted by Ahmet et al. and others, such as the use of a clamp device lowered the efficacy rate of ablation and resulted in a longer follow-up duration, as shown in Table 1.

Some limitations of this study should be considered. First, although the selected studies are all RCT studies comparing the effects of TSA and CA in AF patients, several variables were observed in the enrolled studies, such as randomly assigned patterns, the ablation strategies, the definition of AF recurrence and the methods of follow-up. Second, several studies have reported the use of convergent epicardial and endocardial ablation procedure (CVP) for AF recently [49–52], which provided us with another choice of an invasive treatment for AF. When compared to the results of the studies involved in the present research, it was found that CVP may be able to achieve less arrhythmia recurrence with more periprocedural complications in patients with AF. Therefore, based on available data, it is difficult to draw definitive conclusions. This poses as an important limitation of the present study, in which only TSA and CA were compared, but other invasive treatment methods for AF, especially convergent ablation procedures, were not included.

Another limitation of our research is that the sample size was relatively small. As such, the present study only stands to provide a starting point for further studies. Further investigation must be conducted and should include the collection of data from a larger population size. Most of the studies had a follow-up period of 1 year [9, 13, 14, 37, 38], with only one lasting 2 years [10]; the results maybe therefore, only be valid for 1 year, as, if the studies were carried out for a longer period (>1 year), the results of the freedom from AF may be different. Several RCTs are currently underway, which may address the aforementioned limitations [25, 29, 31].

CONCLUSION

Doctors should recommend either TSA or CA to patients with AF after careful consideration of patient factors and the pros and cons of each procedure, as, compared to CA, TSA showed a higher AF ablation efficacy with a higher rate of SAEs. In addition, TSA did not show better efficacy in the sub-analysis of patients with PAF or early PersAF as the first invasive procedure.

SUPPLEMENTARY MATERIAL

Supplementary material is available at ICVTS online.

Funding

This study was sponsored by the Natural Science Foundation of Shandong Province of China (ZR2019PH036) and the Key Research and Development Plan of Jinan Shandong Province of China (201805056).

Conflict of interest: none declared.

Author contributions

Shaolei Yi: Data curation; formal analysis; investigation; methodology; writing—original draft. Xiaojun Liu: Data curation; formal analysis. Wei Wang: Writing—original draft. Lianghua Chen: Data curation; methodology; writing—original draft; writing—review & editing. Haitao Yuan: Data curation; formal analysis; writing—original draft; writing—review & editing.

Reviewer information

Interactive CardioVascular and Thoracic Surgery thanks George Krasopoulos, Richard B. Schuessler, Piotr Suwalski and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

Abbreviations

 
• AF

  Atrial fibrillation

 
• CA

  Catheter ablation

 
• CIs

  Confidence intervals

 
• MD

  Mean difference

 
• OR

  Odds ratio

 
• PersAF

  Persistent AF

 
• PVI

  Pulmonary vein isolation

 
• RCTs

  Randomized controlled trials

 
• SAE

  Severe adverse event

 
• TSA

  Thoracoscopic surgical ablation