Ablation versus medical therapy for patients with atrial fibrillation: An updated meta‐analysis

Abstract To investigate the effect of ablation compared to medical therapy on clinical outcomes of patients with atrial fibrillation (AF). PubMed, Scopus, Embase, and Web of Science databases were searched using ablation, medical treatment, AF, and related words. The effect of ablation and medical therapy was sought to be gathered on stroke or transitional ischemic attack, mortality, hospitalization, recurrence of AF, progression of AF, and left ventricular ejection fraction. Analyses were performed using R software. 31 studies (the results of 27 randomized controlled trials), compromising an overall 6965 patients (Ablation, n = 3643; Medical treatment, n = 3322) were reviewed in our study, revealed that catheter ablation would result in substantial benefits for patients with AF without significant difference in serious adverse events compared to medical management (Risk Ratio: 0.92, [95% Confidence Interval (CI), 0.64−1.33]). Catheter ablation in patients with AF significantly resulted in a 29% reduction in all‐cause mortality (RR: 0.71, [95% CI, 0.57−0.88]), a 57% reduction in hospitalization (RR: 0.43, [95% CI, 0.27−0.67]), a 53% reduction in AF recurrence (RR: 0.47, [95% CI, 0.36−0.61]), and a dramatic reduction, 89%, in progression of paroxysmal to persistent AF (RR: 0.11, [95% CI, 0.02−0.65]); also associated with a remarkable improvement in their left ventricular ejection fraction (LVEF) (Mean Difference, MD: 6.84%, [95% CI, 3.27−10.42]) compared to medical therapy. Our study showed that ablation may be superior to medical therapy in patients with AF regarding AF recurrence, mortality, LVEF improvement, hospitalization, and AF progression outcomes.


| INTRODUCTION
Atrial fibrillation (AF) is a supraventricular tachyarrhythmia caused by chaotic atrial activation resulting in ineffective atrial contraction which is known to be the most prevalent sustained cardiac arrhythmia in adults. 1 Medical problems caused by or related to AF can increase morbidity and mortality rate affecting patients' quality of life and life expectancy.Heart failure (HF), stroke, peripheral thromboembolism, cardiomyopathy, renal failure, myocardial infarction, dementia, and death are seen to be related to AF. 1,2 For such problems caused by AF, management of this disease is one of the medical concerns.Rate control achievement with betablockers and calcium channel blockers can help reduce patients' symptoms and prevent adverse cardiovascular outcomes such as HF and cardiomyopathy induced by tachycardia. 3,4Rhythm control plays a significant role in AF outcomes as it was seen that AF progression was significantly lower in patients with obtained sinus rhythm.6][7][8][9] In a recent randomized controlled clinical trial (RCT) conducted to find if there is a superiority of cryo-balloon ablation over antiarrhythmic drugs (AADs) as initial therapy for symptomatic paroxysmal AF, was seen that AF recurrence was significantly lower. 5,8,10However, in another RCT which used radiofrequency ablation instead of cryo-balloon, there was no significant difference between the treated groups who received ablation and AADs as initial therapy. 11Consequently, the best treatment is still controversial.Some meta-analyses have been done to compare outcome of patients with AF, whether treated by catheter ablation or medical therapy, and the results showed some inconstancy.Evaluation of previous meta-analysis showed that most recent and very large trials 10,12,13 have not been included in some of them. 14,15In another recent meta-analysis, progression of AF was not evaluated between medical treatment and ablation, or the analyzed data were inconsistent. 16Furthermore, the results from some others were contradictory.For example, two different previous meta-analysis done by Mao et al. and Asad et al. resulted in nonsignificant stroke/ TIA risk reduction in CA group. 14,15On the contrary, Song et al. in their new meta-analysis resulted that the effect of CA on risk of stroke/TIA in AF patients could assumed to be significant (RR: 0.61, [0.39−0.96]). 16Considering these controversies and inconsistencies, performing a new meta-analysis that covers all aspects of ablation vs medical treatment for treatment of patients with AF and includes all new studies in this field, evaluates all the subgroups and outcomes simultaneously in one study seemed to be essential.Here, we aimed to perform a comprehensive meta-analysis not only to address these conflicts but also to update the current evidence and analysis of subgroups of patients to have a better insight into the effects of ablation and medical therapy for patients with AF especially by considering various subgroups including cryo-balloon or radiofrequency ablation techniques.Also in this comprehensive meta-analysis, for the first time Progression of AF was evaluated as one of the secondary end points of comparing these two different managements.

| METHODS
This study has been conducted according to Preferred Reporting Items for Systematic Reviews (PRISMA) 2020.

| Eligibility criteria
RCTs that compared catheter ablation (radiofrequency or cryosurgery) with medical treatment for patients with AF were considered eligible for review.Exclusion criteria were case series, editorials, erratum, letters to the editor, narrative reviews, conference abstracts, and retracted articles.

| Information sources and search strategy
PubMed, Scopus, Embase, and Web of Science databases were searched on December 24th of 2021 using a combination of the words "Radiofrequency Ablation," Cryosurgery," "Catheter Ablation," "Atrial Fibrillation," "Best medical treatment," and their related synonyms.The title/abstract filter was the only filter that was used, and no automated search tool was used for the search process.The comprehensive search strategy could be found in the Supporting Information Material.

| Selection process and data collection process
The yield of our search was exported into an endnote library (Version X 8, Clarivate Analytics) and then duplicates were omitted manually.
The screening process was started using the Rayyan web-based tool. 17For screening two separate authors (A. S. and F. K.) were involved and conflicts were instigated by a third author (A.A.) at that application.In this step of screening, titles and abstracts were investigated for meeting our inclusion and exclusion criteria.In the second step, included articles were exported from the Rayyan app into an endnote library and full texts were added and screened.

| Data extraction
An excel spread sheet has been created to extract the outcomes of the eligible articles.The desirable outcomes were stroke or transitional ischemic attack, mortality, hospitalization, recurrence of AF or any atrial tachyarrhythmia (AT), progression of AF all as events, and change in the mean of left ventricular ejection fraction (LVEF).
The data extraction was first performed by one author and then rechecked by two other researchers.The effect of the intervention in patients with HF, paroxysmal, persistent, naive, or refractory AF and different types of catheter ablation like cryo-balloon and radiofrequency were subsets that were sought to be gathered as subgroups.

| Risk of bias assessment
The risk of bias for RCTs was assessed using RevMan (version 5.4, Cochrane) 18 and reported as two separate figures.This tool classified each domain into high risk, low risk, and unclear risk of bias.

| Synthesis methods
The core analysis was carried out using R software (Version 4). 19A Random-effect model was considered for analysis.Heterogeneity was assessed using I 2 , and heterogeneity of more than 50% was considered significant.Pooled data and subgroups were reported in a forest plot.Publication bias was assessed using funnel plots and quantitively using Egger's test (for continuous data) and Petter's test (for binary data) for analyses that included nine and more studies.A p Value of less than .05 in both tests was considered as a significant publication bias and a trim and fill test was performed to consolidate the results in cases that any significant publication bias was detected.
Sensitivity analysis was performed using the leave-one-out method.

| Study selection
Our search emerged in 17 747 articles and after removing duplications, 9387 articles remained for the two-step screening process.The screening process ensued in 31 articles selected for our current study for further evaluation.Finally, 31 articles (from 27 different RCTs) were incorporated in this systematic review.
PRISMA flowchart clarifies the comprehensive study selection process (Figure 1).
The ablation method in the majority of our selected studies was radiofrequency catheter ablation, except for one study using hot balloon ablation, 26 and four other studies using cryo-balloon ablation. 5,8,10,37The comparator medical therapy group in most of the selected studies used AADs, while four trials used the rate control medications, 29,30,32,35 and three trials used a combination of both. 7,31,38

| Risk of bias in studies
Quality and risk of bias assessment have been illustrated in Supporting Information S1: Figure 1 in the Supporting Information Material.

| Stroke/TIA
Analysis of 12 studies for determining the overall effect of the two interventions on stroke and/or TIA revealed a nonsignificant reduction (RR: 0.71 [95% CI, 0.44−1.14],I 2 : 0%, p = .57).The interventions' effect on stroke/TIA only in studies with a minimum follow-up of 12 months revealed the same result (RR: 0.62, [0.37−1.04],]42 (Forest plots of Stroke/TIA outcomes have been illustrated in Supporting Information S1: Figures 6−9 in the Supporting Information Material).

| Change in left ventricular ejection fraction
Analysis of different studies showed that comparing Ablation versus BMT in the management of AF showed that catheter ablation also resulted in a remarkable greater improvement in LVEF in a follow-up range of 6 to 60 months (mean difference, MD: 6.84, [95% CI; 3.27−10.42],2]34,35 Improving LVEF especially was beneficial for persistent AF patients with concordant HFrEF (MD: 6.39, [2.26−10.53],I 2 : 82%, p < .01).
F I G U R E 2 Forest plot of studies comparing medical therapy with two different methods of ablation in term of SAE in AF patients (radio-frequency catheter ablation vs. BMT/cryo-balloon catheter ablation vs. BMT).AF, atrial fibrillation; SAE, serious adverse events.
(Forest plots of the LVEF improvement outcome have been illustrated in Figure 5A, and Supporting Information S1: Figure 18 in the Supporting Information Material).

| AF progression
Analysis of 3 RCTs revealed that ablation is the superior management for paroxysmal AF compared to medical therapy, as it significantly delays the progression of the condition to persistent AF with a RR reduction of 89% (RR: 0.11, [0.02−0.65],I 2 : 0%, p = .49) 21,24,25igure 5B).

| Publication bias
Publication bias was only significant for the analysis of the investigation of the effect of ablation on AF recurrence events

| DISCUSSION
To the best of our knowledge, this is the largest meta-analysis that compared the effect of ablation with medical treatment in patients with AF and is the only one addressing the effect of cryo-balloon ablation in the subgroup analyses.Analyses revealed that AF ablation results in a significantly lower mortality rate, hospitalization, AF recurrence, or any atrial arrhythmia recurrence and a further improvement in LVEF compared to medical management.These effects were both valid using cryo-balloon and radiofrequency ablation methods while the risk of serious side effects is not increased significantly.Data on cardiovascular hospitalization and HF hospitalization was not adequate for meta-analysis.
Our study showed that ablation would result in a significantly greater reduction in AF or any AA recurrence in patients with refractory, paroxysmal, or persistent AF.It is probable that this decrement in atrial arrhythmia recurrence would result in better patients' quality of life and exercise capacity tolerance, lower morbidity and complications.This study revealed that ablation would cause a remarkable 47 percent risk reduction of atrial arrhythmia recurrence in patients with PAF who were naïve to treatment.
The significant beneficial effect of ablation on recurrence of AA in the naïve subgroup in our study was in agreement with Andrade et al. study that revealed the recurrence of atrial tacky arrhythmia was significantly lower in the cryo-balloon ablation group (42.9% vs. 67.8%,HR: 0.48 [95% CI, 0.35−0.66];p < .001).It also showed that the effect of cryo-balloon ablation on decreasing the recurrence of symptomatic atrial arrhythmia was even more (11% vs. 26.2%,HR: 0.39 [95% CI, 0.22−0.68]). 5The RAAFT-2 study also demonstrated that AF patients who underwent radiofrequency ablation as first-line treatment would less likely to experience the recurrence of symptomatic atrial arrhythmia compared to AAD therapy (HR, 0.56 [95% CI, 0.33−0.95];p = .03)also the overall number of atrial arrhythmia episodes were significantly lower in ablation group (HR, 0.33 [95% CI, 0.28−0.4],p < .001). 22Seemingly, Poole et al. selected 56% of the CABANA population to study atrial arrhythmias recurrence utilizing a Trans-Telephonic Monitor electrocardiogram monitoring system and demonstrated that ablation was effective in reducing any AF recurrence (by 48%) and reducing symptomatic AF (by 51%) in a total 60 months follow-up. 13The study revealed that ablation significantly reduces the recurrence of symptomatic AF (HR: 0.49, [0.39−0.61]),time to the first recurrence of any AF (0.52, [0.45−0.6]),and time to first atrial arrhythmia recurrence (0.53, [0.46−0.62]) 13lation also might cause a relative risk reduction of stroke/TIA compared to medical management as it could result in a better sinus rhythm maintenance and therefore lower risk of thrombosis due to lesser blood stasis in LA.The result of our study showed that the risk reduction of stroke/TIA was not significant in the ablation group.There were some confounding factors including assuming procedural TIAs without any significant complication with the same value of disabling stroke events, and another factor might be the inadequate follow-up time in some of the trials for assessing long-term AF complications like stroke or TIA.We could not eliminate the first factor due to lack of available information, but for minimizing the latter mentioned confounding factor, we decided to check the interventions' effect on stroke/TIA only in studies with a minimum follow-up of 12 months, but again the result was not significant.
Likewise, two different previous meta-analyses done by Mao et   14,15 In contrast, Song et al. in their metaanalysis declared that the effect of ablation on the risk of stroke/TIA in AF patients could be assumed to be significant (RR: 0.61, [0.39−0.96])These results might be due to separating AF induced stroke from procedural ones. 16alyses also showed that CA would result in a remarkable decrease in total hospitalization rate of AF patients compared to medical management but we could not have statistical analysis on cardiovascular or HF hospitalization (CVH or HFH), whereas some of our included studies showed a significant reduction in CVH, and HFH. 23,24In agreement to our findings, CASTLE-AF trial showed a significant lower HFH (HR: 0.56 [0.37−0.83])and CVH (HR: 0.72 [0.52−0.99])and even lesser cardiovascular death (0.49 [0.29−0.84]). 7CABANA also showed that death and CV hospitalization were somehow lower in the ablation group compared to medical therapy (HR: 0.83, [0.74−0.93],p = .001). 38These estimated effects of ablation had been influenced by treatment crossover (27.5% from BMT to ablation) and lower event rates than anticipated. 38e most considerable decrease in the rate of hospitalization in statistical analyses was noticed in the naive AF subgroup who has not received any treatment for AF before the trial.It could be a sign of better morbidity control and complication reduction by considering ablation as the first-line treatment in PAF patients who are naive to therapy.
Our analyses for the first time suggested that compared to medical therapy, ablation could significantly reduce and delay the progression to persistent AF, so patients with AF might take benefit from the consideration of early ablation.Seemingly, the ATTEST study revealed that early ablation could be the superior management for delaying the progression of the disease, as the development of persistent AF/AT in the ablation group was 10 times lower (HR: 0.107 [0.024−0.47];p = .0031)with a considerably lower rate of persistent AF (0.0% vs. 10.2%;p = .0002). 21[45] The ablation of AF in patients with concomitant HF seems too very beneficial.AF and HF could form a vicious cycle as persistent tachycardia due to AF may result in arrhythmia-induced cardiomyopathy, [46][47][48] and HF is a risk factor for AF development and The study also studied the results of patients in different subgroups according to their LVEF and showed that the effect of ablation on reduction in mortality was especially greater in patients with LVEF of 50% or more (HR: 0. This study has faced some limitations.Due to a lack of data on the distinction between noncomplicated procedural TIA and disabling stroke, we could not separate procedural TIA from stroke/TIA resulting from AF thromboembolic events.We tried analyses of studies with follow-up of at least 12 months to minimize this confounding factor, but we could not remove these confounding factors' effect on our results.Lack of data on some of our subgroups made it impossible to do all the subgroup analyses completely.We were supposed to have another subgroup analysis on comparing the effect of crayon-balloon ablation versus radiofrequency catheter ablation as two different methods of ablation in AF patients and see which one is more effective, but as cryo-balloon was used just in trials assessing AF patients who were naïve to AAD therapy and RF CA population was not the same (refractory AF to AAD, persistent AF, PAF with HF), the combination and comparing these two subgroups became unreasonable and impossible.Publication bias tests resulted positive in the analysis of AF recurrence, it could be due to the high-quality and legitimate included RCTs. 50 : 0.44, [0.29−0.66],I 2 : 63%, p = .04)(RF CA vs. BMT, RR: 0.46, [0.38−0.56],I 2 : 85%, p < .01).

(
Forest plots of the hospitalization outcome have been illustrated in Supporting Information S1: Figures 14−17 in the Supporting Information Material).

F I G U R E 3
Forest plot of included studies investigating the effect of ablation on atrial arrhythmia recurrence compared to medical treatment in several subgroups.(A: Comparing common methods of ablation, B−E: subgroup analysis in different atrial fibrillation [AF] subtype).(A) Cryo-balloon and radiofrequency catheter ablation.(B) Naïve AF (did not receive any treatment for their AF condition before the trial).(C) Paroxysmal AF. (D) Refractory AF (patients with AF who had failed at least one AAD before trial).(E) Persistent AF.F I G U R E 4 Forest plot of included studies investigating the effect of ablation on hospitalization compared to medical treatment.F I G U R E 5 (A) Forest plot of studies comparing Ablation versus medical therapy in term of LVEF improvement in AF patients.(B) Forest plot of studies comparing Ablation versus medical therapy in term of progression of paroxysmal AF to persistent AF.AF, atrial fibrillation; LVEF, left ventricular ejection fraction.compared to medical treatment.A Trim and fill test was carried out and resulted in an adjusted RR of 0.652 with a 95% CI of 0.455−0.934.
Characteristics of the included randomized controlled trials.