Meta-Analysis of Randomized Controlled Trials of Atrial Fibrillation Ablation With Pulmonary Vein Isolation Versus Without

Objectives This meta-analysis examined the ability of pulmonary vein isolation (PVI) to prevent atrial fibrillation in randomized controlled trials (RCTs) in which the patients not receiving PVI nevertheless underwent a procedure. Background PVI is a commonly used procedure for the treatment of atrial fibrillation (AF), and its efficacy has usually been judged against therapy with anti-arrhythmic drugs in open-label trials. There have been several RCTs of AF ablation in which both arms received an ablation, but the difference between the treatment arms was inclusion or omission of PVI. These trials of an ablation strategy with PVI versus an ablation strategy without PVI may provide a more rigorous method for evaluating the efficacy of PVI. Methods Medline and Cochrane databases were searched for RCTs comparing ablation including PVI with ablation excluding PVI. The primary efficacy endpoint was freedom from atrial fibrillation (AF) and atrial tachycardia at 12 months. A random-effects meta-analysis was performed using the restricted maximum likelihood estimator. Results Overall, 6 studies (n = 610) met inclusion criteria. AF recurrence was significantly lower with an ablation including PVI than an ablation without PVI (RR: 0.54; 95% confidence interval [CI]: 0.33 to 0.89; p = 0.0147; I2 = 79.7%). Neither the type of AF (p = 0.48) nor the type of non-PVI ablation (p = 0.21) was a significant moderator of the effect size. In 3 trials the non-PVI ablation procedure was performed in both arms, whereas PVI was performed in only 1 arm. In these studies, AF recurrence was significantly lower when PVI was included (RR: 0.32; 95% CI: 0.14 to 0.73; p = 0.007, I2 78%). Conclusions In RCTs where both arms received an ablation, and therefore an expectation amongst patients and doctors of benefit, being randomized to PVI had a striking effect, reducing AF recurrence by a half.

I n a randomized controlled trial (RCT), when a patient is randomized to an arm without a procedural intervention instead of an arm with a procedural intervention, both the patient and the patient's medical care staff are inevitably aware that the patient has not had a procedure which other patients have had. This can easily lead to a lower threshold for reporting concerns and therefore undergoing followup tests. This in turn can lead to more adverse events becoming documented simply through the increased vigilance and opportunity for detection. This phenomenon has been termed "subtraction anxiety" (1).
Blinding the patient to treatment allocation removes this phenomenon, but this can be challenging when invasive procedures are involved.
Pulmonary vein isolation (PVI) is the cornerstone of atrial fibrillation (AF) ablation procedures. However, its efficacy has been formally judged only compared to medical therapy in open-label trials (2).
There is, however, a set of trials of AF ablation that allow us to establish the efficacy of pulmonary vein isolation relative to other methods of AF ablation.
These trials have the additional advantage that, in the comparison arm where alternative ablation takes place, the patient has received a procedure and therefore the patients (and the doctors downstream) do not automatically feel subtraction anxiety. A particularly informative subset of these trials are those in which both arms received the same ablation procedure, except that one arm had PVI and the other did not. Those trials are informative of the incremental effect of PVI.
The present meta-analysis examined the ability of PVI to prevent AF in RCTs in which the patients who were not receiving PVI nevertheless underwent a procedure.

METHODS
This study carried out a meta-analysis of RCTs evaluating AF ablation, comparing a strategy involving PVI with a strategy not involving PVI.  Analysis was based on the intention-to-treat outcomes from each study. The number of patients at risk, the total patients in the arm, and the number lost to follow-up were extracted from each arm. Data were also extracted regarding symptom scores where present. Where these data were not presented, they were calculated as described in the Online Appendix. A random-effects meta-analysis was performed using the restricted maximum likelihood estimator.
Moderator variables were assessed using a mixedeffects meta-analytical model. The I 2 statistic was used to assess heterogeneity. R software (3) with the Metafor feature (4) was used for all statistical analysis. The Cochrane risk-of-bias tool was used to assess included studies (5). Tests for publication bias were not performed because <10 trials were included for analysis (6). The Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline was used to report results (7). Values are mean AE SD, unless otherwise stated. Sensitivity analyses were performed, excluding each trial in turn.

MEDLINE and
Cochrane database searches yielded a total of 2,544 studies (Online Figure 1). Seven studies met the inclusion criteria, 1 study was subsequently excluded. Therefore, 6 studies formed the final analysis and are detailed in Table 1 and Online Table 2. Sau et al.
A U G U S T 2 0 1 9 : 9 6 8 -7 6 Pulmonary Vein Isolation vs. Alternative Ablation Targets The Cochrane risk-of-bias tool was used to assess trial quality (Online Table 1, Online Figure 2 (Figure 3).    (13), renal denervation for hypertension (14), and patent foramen ovale (PFO) closure for migraine (15 anxiety. The merit of these 6 trials is that each of them had an invasive procedure in both arms. All patients therefore believed that they had undergone a  to an overestimation of the efficacy of PVI in this analysis.

FIGURE 3 Effect of PVI Versus Non-PVI Ablation on AF Recurrence Stratified by the Control Arm Ablation Strategy
The efficacy of complete versus selective PVI has been studied (22). That analysis suggested that isolation of arrhythmogenic PVs alone is comparable to empirical isolation of all PVs. As the aim of both arms was to isolate pulmonary vein triggers for AF, it was not included in this analysis.

CONCLUSIONS
Analysis of the present study provides an estimate of the efficacy of PVI in preventing AF/AT. This is a reduction of one-half and perhaps even two-thirds if only the purest trials are considered. A true placebo- TRANSLATIONAL OUTLOOK: AF ablation has not undergone a comparison trial relative to an invasive placebo control. The correspondence between arrhythmia and symptoms is imperfect; therefore, it is unclear how reductions in AF burden, which PVI appears to do effectively, relates to symptom regression. A placebo (sham)-controlled RCT would be the ideal method for testing this.