Efficacy and safety of dronedarone in patients with a prior ablation for atrial fibrillation/flutter: Insights from the ATHENA study

Abstract Background The role of antiarrhythmic drugs for atrial fibrillation/atrial flutter (AF/AFL) after catheter ablation is not well established. Hypothesis We hypothesized that changing the myocardial substrate by ablation may alter the responsiveness to dronedarone. Methods We assessed the efficacy and safety of dronedarone in the treatment of paroxysmal/persistent atrial fibrillation/atrial flutter (AF/AFL) post‐ablation, based on a post hoc analysis of the ATHENA study. A total of 196 patients (dronedarone 90, placebo 106) had an ablation for AF/AFL before study entry. In these patients, the effect of treatment on the first hospitalization because of cardiovascular (CV) events/all‐cause death was assessed, as was AF/AFL recurrence in individuals with sinus rhythm at baseline. The safety of dronedarone vs placebo was also determined. Results In patients with prior ablation, dronedarone reduced the risk of AF/AFL recurrence (hazard ratio [HR]: 0.65 [95% confidence interval [CI]: 0.42, 1.00]; P < .05) as well as the median time to first AF/AFL recurrence (561 vs 180 days) compared with placebo. The HR for first CV hospitalization/all‐cause death with dronedarone vs placebo was 0.98 (95% CI: 0.62, 1.53; P = .91). Rates of treatment‐emergent adverse events were 83.1% vs 75.5% and rates of serious TEAEs were 27.0% vs 18.9% in the dronedarone and placebo groups, respectively. One death occurred with dronedarone (not treatment‐emergent) and five occurred with placebo. Conclusion In patients with prior ablation for AF/AFL, dronedarone reduced the risk of AF/AFL recurrence compared with placebo, but not the risk of first CV hospitalization/all‐cause death. Safety outcomes were consistent with those of the overall ATHENA study.


| INTRODUCTION
Antiarrhythmic drugs (AADs) have commonly been used as first-line treatment for maintenance of sinus rhythm in individuals with paroxysmal and persistent atrial fibrillation (AF), and form a class IA recommendation in the American Heart Association (AHA)/American College of Cardiology (ACC)/Heart Rhythm Society (HRS) AF guidelines. 1,2 Real-world data indicate an increasing use of catheter ablation for rhythm management of AF, 3,4 driven by improvements in ablation techniques and outcomes. 5 6,7 The use of AADs and catheter ablation are not mutually exclusive. 1,2,5 AADs are often used immediately following ablation to reduce the risk of early AF recurrence, [8][9][10] and may be used in the long term depending on patient and physician preference. 5 Successful ablation may alter the responsiveness to AADs, 11 possibly as a result of changes to the substrate within the atrium and the triggers for AF. However, there is a substantial gap in scientific evidence regarding the impact of ablation on the effectiveness and safety of AADs in the post-ablation setting.
Dronedarone is an AAD that has been shown to reduce the rate of hospitalization because of cardiovascular (CV) events in patients with paroxysmal or persistent AF or atrial flutter (AFL) in a randomized, double-blind, placebo-controlled phase 3 study ("A Trial with Dronedarone to Prevent Hospitalization or Death in Patients with Atrial Fibrillation" [ATHENA; NCT00174785]). 12 The aim of the current analysis was to assess the efficacy and safety of dronedarone among patients with prior ablation who received treatment in the ATHENA study. Patients were randomized in a 1:1 ratio to dronedarone 400 mg twice daily or placebo in addition to their rate-control therapy; the minimum follow-up period was 12 months. The primary efficacy endpoint was first hospitalization because of CV events or death from any cause. Secondary endpoints included death from any cause, CV death and first CV hospitalization. The study was approved by the independent review board at each participating site and was con-

| Post hoc analysis
This analysis focused on patients who had received any ablation procedure for AF/AFL before randomization in the ATHENA study and subsequently received study treatment. The primary and secondary endpoints of the ATHENA study were assessed in this patient population. First recurrence of AF/AFL was assessed in patients who were in sinus rhythm at baseline. In addition, the effect of ablation status on the incidence of treatment-emergent adverse events (TEAEs) occurring from first study drug intake up to last study drug intake plus 10 days was assessed.

| Assessments
Clinical evaluations, categorization for unplanned hospitalizations (CV or non-CV), deaths (non-arrhythmic cardiac, arrhythmic cardiac, non-cardiac vascular, or non-CV) and safety assessments were conducted as described previously. 12 Scheduled 12-lead ECGs were recorded at 7 and 14 days, and at 1, 3, and 6 months after randomization, then every 6 months thereafter. In addition, ECGs were recorded during unscheduled visits, such as in patients presenting with recurrent symptoms. Each ECG was classified by the investigator as demonstrating AF, AFL or sinus rhythm. AF/AFL recurrence was determined to occur at the first instance of cardioversion, AF/AFL based on electrocardiography or hospitalization for AF/AFL, as described previously. 14 Patient baseline and demographic characteristics were reviewed for patients with and without ablation before study randomization.

| Statistical analysis
For the primary endpoint, cumulative incidence functions in each treatment group were calculated and plotted using the Kaplan-Meier method. Unadjusted hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using an unstratified Cox regression model with treatment group as the only factor. A multivariate analysis was considered unsuitable based on the small patient population and multiple biases associated with a post hoc analysis. Data were analyzed using SAS version 8.2 or later (Cary, North Carolina). Potential interaction between treatment (ie, dronedarone or placebo) and ablation status (ie, prior ablation or no prior ablation) was assessed using a Cox regression model for occurrence of first TEAE applied to the whole study safety population. An independent data and safety monitoring board provided regular reviews of study data.

| Baseline and demographic characteristics
Among the 4628 patients randomized in the ATHENA study, 217 (4.7%) had received any ablation procedure before randomization, including 196 (4.2%) who had an ablation for AF/AFL. The proportion of patients who received ablation for AF/AFL before randomization was similar in the dronedarone (90 of 2301 patients; 3.9%) and placebo (106 of 2327 patients; 4.6%) groups (Table 1).
Patient baseline and demographic characteristics in the dronedarone and placebo groups among patients with and without prior ablation are shown in Table 2. Among patients who had received prior ablation for AF/AFL, baseline and demographic characteristics were similar in the dronedarone and placebo groups.

| Efficacy
Median follow-up for patients who received ablation for AF/AFL was 666 days in those treated with dronedarone and 688 days in those treated with placebo.
In patients with prior ablation, first CV hospitalization or death from any cause occurred in 35 of 90 patients (38.9%) randomized to dronedarone and in 42 of 106 patients (39.6%) randomized to placebo  Table 3).
The most common cause of first CV hospitalization was AF/supraventricular rhythm disorders, accounting for 37% of first CV hospitalizations with dronedarone and 46% with placebo (Table S1).
One death (because of cardiogenic shock) occurred in the dronedarone group and was not treatment-emergent; five deaths (two sudden cardiac deaths, and one case each because of pneumonia, pancreatic adenocarcinoma, and multi-organ failure) occurred in the placebo group.   Figure 1).

| Safety
TEAEs occurred in 83.1% and 75.5% of patients treated with dronedarone and placebo, respectively ( Table 4). The most common TEAEs in the dronedarone and placebo groups were dizziness (12.4%

| DISCUSSION
In this analysis, dronedarone treatment decreased the risk of first AF/AFL recurrence compared with placebo in patients who had received catheter ablation before randomization, and demonstrated a safety profile similar to that observed in the overall ATHENA population. 12 The rate of AF/AFL recurrence was 57% in the dronedarone group vs 71% in those treated with placebo.
The rates of first CV hospitalization or death from any cause (39% in the dronedarone group and 40% in the placebo group) in patients with prior ablation were in fact similar to those observed in the placebo arm of the overall ATHENA population (39%), 12   In real-world evaluations, fewer than 50% of patients achieve treatment success at 1 year following catheter ablation, based on freedom from AF recurrence without the use of AADs. 15,16 Data from the ESC-EHRA AF ablation long-term registry show that for patients receiving ablation, 90% had received AAD treatment before ablation, 68% received AADs shortly after ablation and 46% were on AADs at 1-year follow-up. 17 In the CABANA study, a high incidence of crossovers between the ablation and medical therapy arms highlights the difficulty in separating ablation and medical therapies for AF rhythm management. 7 Furthermore, almost 50% of patients in the ablation arm of the CABANA study had AF recurrences at 4 years, emphasizing the need to explore efficacy and safety of AAD treatment as complementary rather than competing approaches to rhythm management.
While short-term adjunctive AAD therapy has been shown to significantly reduce the risk of early AF recurrence compared with ablation alone, 9 there are few randomized controlled trials and real-world data addressing outcomes with AADs in patients with prior ablation, and the efficacy and safety of AADs in this setting are unclear. 5,18,19 Data from a large, single-center study in 439 patients with paroxysmal or persistent AF suggest that shortened time to recurrence after ablation is linked to an increased responsiveness to AADs, possibly as a result of substrate changes, making reintroduction of AAD therapy after ablation a compelling option. 11

| Limitations
The ATHENA study was not designed or powered to detect differences in efficacy or safety of dronedarone vs placebo within the subset of patients with prior ablation. As a post hoc analysis, and with a relatively small number of patients, this study should be regarded as exploratory. Furthermore, information was not available on either the timing or characteristics of prior ablation, the type of AF/AFL for individual patients at randomization, or heart rate during sinus rhythm and AF episodes throughout the study.
T A B L E 3 Efficacy of dronedarone vs placebo in patients with ablation for AF/AFL before randomization in the ATHENA study Time period from the first study drug intake up to the last study drug intake plus 10 days ("treatment period"); data include all AEs/laboratory abnormalities occurring in ≥5% of patients in either group plus selected AEs/laboratory abnormalities previously reported for the ITT ATHENA population. 12 b Occurring from first study drug intake up to last study drug intake plus 10 days. c One of these cases of QT prolongation leading to permanent study drug discontinuation occurred before first study drug intake.