Catheter vs thoracoscopic ablation for atrial fibrillation: Meta-analysis of randomized trials.

BACKGROUND: We meta-analyzed the efficacy and safety of catheter (CA) vs thoracoscopic (TA) ablation for atrial fibrillation (AF).
METHODS: PubMed, Embase, and Cochrane databases were searched until 31/12/2019 for relevant randomized trials and subsequent pooled analyses.
RESULTS: In six trials totaling 465 patients (254 CA/211 TA), 1-year AF recurrences were higher for CA (46% vs 26%, odds ratio 2.90 [95% CI 1.32-6.38]), whereas total adverse events were lower (10% vs 25%, 0.35 [0.14-0.86], respectively).
CONCLUSION: CA has lower efficacy but higher safety than TA. CA should remain the first-line AF ablation strategy, and TA reserved for selected CA-resistant patients where rhythm control is clinically necessary.

INTRODUCTION

As a global epidemic, atrial fibrillation (AF) is the commonest form of arrhythmia and a well‐established risk factor of mortality and morbidities such as stroke and heart failure. , The four facets for AF management include lifestyle modification, anticoagulation, rate, and rhythm control. , Antiarrhythmics and ablation procedures are strategies for rhythm control, and recent trials have shown potential benefits of catheter ablation to reducing cardiovascular events and death in AF patients especially with heart failure. , Catheter (CA) and thoracoscopic (TA) ablation are the two approaches for AF ablation, however, the optimal approach remains controversial because of the paucity of randomized trials. This meta‐analysis aimed to compare the efficacy and safety outcomes for CA vs TA at treating AF across randomized trials.

METHODS

We searched Pubmed, Embase, and Cochrane databases from 1 January 1980 to 31 December 2019 using search terms “atrial fibrillation,” “ablation” or “pulmonary vein isolation,” “catheter or percutaneous,” and “minimally invasive or thoracoscopy.” Eligible studies must be original randomized trials reporting both CA and TA outcomes. Two authors screened studies for inclusion, and data pertaining to study design, characteristics, efficacy, and adverse events were recorded. Reviewer Manager (Version 5.3.5, The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) program was used for meta‐analyses using random effects models. Pooled odds ratios (OR) with 95% confidence intervals (95% CI) were calculated and Forrest Plots constructed. The primary efficacy endpoint was AF recurrence at 1‐year, and primary safety endpoint was total adverse events. Cochrane Q and I2 statistic were used to assess studies' heterogeneity, and Funnel plots for publication bias. The PRISMA guidelines were followed during the conduct of this meta‐analysis.

RESULTS

There were 652 articles generated from the search, and the abstract screening excluded 122 duplicate studies, 193 unrelated studies, 162 reviews, 148 single arm studies, and eight nonrandomized studies. Amongst 19 full‐text articles reviewed, five reviews, five nonrandomized studies and three incomplete randomized trials were excluded, leaving six randomized trials totaling 465 patients (254 CA and 211 TA) eligible. , The study design and clinical characteristics for included studies are listed in Tables 1 and 2.

TABLE 1

Design of eligible trials of catheter (CA) vs thoracoscopic (TA) ablation

Author	Year	Country	Center	Date	N (CA/TA)	Follow‐up (mo)	AF recurrence detection
Boersma 7	2012	Spain, Netherlands	2	2007/7‐2010/6	63/61	12	7‐day Holter at 12 mo
Pokushalov 8	2013	Russia	1	2011/1‐2011/11	32/32	12	Loop recorder 0‐12 mo
Wang 9	2014	China	1	2008/3‐2012/3	72/66	12	24‐hour Holter at 1, 3, 6, and 12 mo
Adiyaman 10	2018	Netherlands	1	2007/11‐2015/2	27/23	24	7‐day Holter at 12 and 24 mo
Sindby 11	2018	Denmark	1	2011/4‐2014/1	11/10	12	7‐day Holter at 6 and 12 mo
Sugihara 12	2018	United Kingdom	1	2012/4‐2015/8	49/19	12	Loop recorder/pacemaker 0‐12 mo

TABLE 2

Clinical characteristics reported by eligible trials of catheter (CA) vs thoracoscopic (TA) ablation

	Number of studies	CA	TA
N	6	254	211
Demographics
Age (y)	6	57 ± 9	55 ± 8
Male	6	173/254 (68%)	145/211 (69%)
Body mass index (kg/m2)	3	28 ± 4	28 ± 5
Past history
Hypertension	5	80/191 (42%)	60/150 (40%)
Diabetes	5	25/191 (13%)	14/150 (9%)
Stroke	3	7/153 (5%)	10/117 (9%)
Ischemic heart disease	2	10/112 (9%)	2/80 (3%)
Atrial fibrillation
Paroxysmal	6	214/254 (84%)	183/211 (87%)
Duration (years)	4	5.9 ± 5.2	5.9 ± 5.2
Antiarrhythmic use	5	195/205 (95%)	181/192 (94%)
Prior failed catheter ablation	5	78/182 (43%)	81/145 (56%)
Echocardiography
Left ventricular ejection fraction (%)	5	60 ± 7	59 ± 6
Left atrial diameter (mm)	5	44 ± 8	44 ± 8
Procedure
Procedure time (min)	6	144 ± 39	213 ± 68
Fluoroscopy time (min)	4	16 ± 9	0 ± 0
Left atrial appendage ligation	5	0/182 (0%)	144/145 (99%)

Results of the meta‐analysis are shown in Table 3, and Forrest plots for the main efficacy and safety outcomes illustrated in Figure 1. Pooled rates of AF recurrences at 6 months were CA 38% vs TA 19%, OR 2.91, 95% CI 1.64‐5.16 in three studies; AF recurrences at 1‐year were 46% vs 26%, OR 2.90, 95% CI 1.32‐6.38 in six studies; and total adverse events CA 10% vs TA 25%, OR 0.35, 95% CI 0.14‐0.86. CA also had higher pooled rates of repeat ablation during follow‐up, but lower rates of pneumothorax, hemothorax, or hydrothorax complications (all P < .05). There was no significant heterogeneity observed for all outcomes except AF recurrences at 1‐year, and no significant publication bias found (Funnel plots not shown).

TABLE 3

Summary of meta‐analysis findings of catheter (CA) vs thoracoscopic (TA) ablation for atrial fibrillation (AF)

Outcome	Number of studies	CA	TA	Odds ratio	95% confidence interval	P value
Efficacy
AF recurrence at 6 mo	3	38%	19%	2.91	1.64‐5.16	<.001
AF recurrence at 12 mo	6	46%	26%	2.90	1.32‐6.38	.008
Repeat ablation	2	25%	2.0%	10.4	1.85‐59.7	.008
Safety
Total adverse events	5	10%	25%	0.35	0.14‐0.86	.02
Death	5	0.0%	0.7%	0.12	0.00‐3.20	.21
Stroke	4	1.5%	0.8%	1.59	0.19‐13.2	.67
Pacemaker implantation	2	0.0%	1.7%	0.19	0.01‐3.98	.28
Pneumothorax	3	0.0%	7.8%	0.09	0.01‐0.74	.03
Hemothorax and hydrothorax	5	0.0%	5.5%	0.15	0.03‐0.69	.02
Pericardial effusion	3	0.8%	2.6%	0.48	0.08‐2.81	.41

FIGURE 1

Forest plots of pooled outcomes of catheter (CA) vs thoracoscopic (TA) ablation—(A) atrial fibrillation recurrence at 6 mo, (B) atrial fibrillation recurrence at 12 mo, and (C) total adverse events

DISCUSSION

This meta‐analysis has some key findings. First, TA had higher efficacy shown by lower rates AF recurrence and repeat ablation than CA. Second, TA also had higher rates of adverse events particularly thoracic complications. The key clinical characteristics of eligible randomized trials include middle‐aged patients with a high prevalence of paroxysmal AF in 85%, AF duration of 5.9 years, high antiarrhythmic use in 95%, prior failed CA in approximately half, and normal mean left ventricular ejection fraction, but mildly dilated left atrial dimensions on echocardiography. , , , , , Therefore, patients with paroxysmal AF refractory to antiarrhythmic drugs and previous failed CA, but with preserved left ventricular function, are those that our findings best apply to.

The greater efficacy seen in five of six trials for TA , , , , may be explained by TA being more extensive with additional left atrial lines, targeting of epicardial ganglionic plexi and left atrial appendage ligation. Only one study showed no difference in efficacy, in patients without previous CA, suggesting TA should probably not be attempted as a first‐line therapy before CA. Furthermore, the lower rate of repeat ablation during follow‐up (25% vs 2%) is another distinct advantage of TA. However, the high adverse event rates in up to 25% of TA even in randomized trials is important, observed across four of five studies reporting this. , , , In particularly, thoracic complications much higher in TA are detrimental because of their frequent need of invasive procedures, prolonging of intensive care and/or hospital stay, and potential for long‐term sequelae. Two previous meta‐analyses reported similar findings for CA compared to TA, however, both of these included observational studies subjective to confounding bias, and being earlier did not contain the three recent important randomized trials since 2018, so our findings are the most contemporary and least biased. ,

Therefore, although TA has higher efficacy, its adverse event rates strongly cautions and diminishes its role in AF management. The Heart Rhythm Society consensus statement for AF ablation in 2017 and the European Society of Cardiology guidelines 2016 both suggest isolated TA can be performed in paroxysmal or persistent AF patients who failed both antiarrhythmic drugs and one or more CA, with a review of relative efficacy and safety compared to CA and consideration of patient choice by a multidisciplinary AF Heart Team. Local expertise of CA and especially TA with sufficient experience and volume are also important for referral. , Our results incorporating more recent trials consolidate the evidence that support these statements. In the setting of concomitant cardiothoracic surgery, however, TA can be performed at a lower threshold in appropriate AF candidates.

This meta‐analysis has some limitations. Several clinically relevant endpoints were not reported including AF burden, symptoms, heart failure, and cost‐effectiveness. The total sample size was moderate at best, restricting the power to identify significant differences in outcomes especially for rare events. Patient‐level data were not available for in‐depth analysis and identification of predictors of efficacy and safety endpoints. There were some heterogeneity in study design, patient characteristics, procedures, medical therapy, and outcome definitions, and studies remain susceptible to publication bias. Our findings also may not necessarily apply to AF patients with very different characteristics to the studies we included. Follow‐up duration of studies was restricted at 12‐24 months.

In conclusion, TA not only had greater efficacy but also higher adverse event rates than CA, particularly thoracic complications. The safety data provide a compelling case for CA to remain the AF ablation strategy of choice. TA should be reserved only for selected patients with AF refractory to antiarrhythmic medications and multiple failed CA, but with ongoing symptoms requiring rhythm control, and our findings support contemporary AF guidelines in terms of the roles of CA and TA.

CONFLICT OF INTEREST

Authors declare no conflict of interests for this article.