Clinical factors related to successful or unsuccessful cardioversion in the EdoxabaN versus warfarin in subjectS UndeRgoing cardiovErsion of Atrial Fibrillation (ENSURE‐AF) randomized trial

Abstract Background EdoxabaN versus warfarin in subjectS UndeRgoing cardiovErsion of Atrial Fibrillation evaluated use of nonvitamin K antagonist oral anticoagulant edoxaban vs enoxaparin‐warfarin in patients with nonvalvular atrial fibrillation undergoing electrical cardioversion. Hypothesis To assess clinical factors related to successful or unsuccessful cardioversion. To evaluate whether differences in adverse events based on anticoagulation strategy may exist. Methods In this multicenter prospective randomized open‐label blinded end‐point evaluation trial, 2199 patients were randomized to edoxaban 60 mg once daily (30 mg for creatinine clearance 15‐50 mL/min, weight ≤ 60 kg, and/or concomitant use of P‐glycoprotein inhibitor) or enoxaparin‐warfarin. Successful cardioversion was confirmed by 12‐lead electrocardiography‐documented sinus rhythm. Results Cardioversion was successful in 1578 patients; in 355 patients, cardioversion was unsuccessful. Male, high body weight, high body mass index (BMI), coronary artery disease, concomitant aspirin, or prior statins use were more common in patients with unsuccessful cardioversion; international normalized ratio control did not differ by cardioversion success. On multivariate analysis, gender (P < .05), body weight (P = .0196) and BMI (P = .0377) emerged as independent predictors of successful cardioversion. There were no significant differences in primary efficacy (a composite of stroke, systemic embolic event, myocardial infarction, and cardiovascular death during overall study period) regardless of cardioversion success. There were no significant differences in bleeding rates, regardless of cardioversion outcome; notwithstanding low numbers, edoxaban and enoxaparin‐warfarin did not differ. Conclusions Male gender, higher mean weight and higher mean BMI were associated with unsuccessful cardioversion. Efficacy and safety outcomes were low and did not differ by cardioversion success.


| ME THODS
The design and principal results of the ENSURE-AF trial (NCT 02072434) have been published. 2 and were treated with warfarin alone; hence, edoxaban was compared to "optimized anticoagulation" with enoxaparin-warfarin. 2 In the TEE-guided subgroup, TEE and ECV had to be performed within 3 days of randomization. In case of presence of thrombi on TEE, patients had an opportunity of completing 28 days of study medication without cardioversion or being discontinued from the study.
The primary efficacy endpoint defined in the main analysis was the composite of stroke, systemic embolic event, myocardial infarction, and cardiovascular death during the overall treatment period from randomization until end of study (28 days on study drug after cardioversion + 30 days follow-up) and the primary safety endpoint was the composite of major and clinically relevant nonmajor bleeding during the on-treatment period (time of first dose to last dose of study drug taken + 30 days). The trial protocol was approved by the ethics committees or institutional review boards in accordance with the Declaration of Helsinki. All patients provided written informed consent prior to participation in the study.
This post hoc analysis focuses on clinical factors related to successful or unsuccessful cardioversion in a contemporary prospective multicenter clinical trial; and reports thromboembolic and bleeding risks on anticoagulation in relation to successful or unsuccessful cardioversion.
Patients were followed for 28 days on study drug after cardioversion plus another 30 days to assess safety, which was
Baseline characteristics of patients are presented in Table 1. During the treatment period, beta blockers were the most commonly used rate control drugs in edoxaban-treated patients (81.1%; n = 865) and in enoxaparin-warfarin-treated patients (81.4%, n = 881). Amiodarone was the most widely prescribed rhythm control agent, in both in edoxaban (41.0%, n = 438) and enoxaparin-warfarin (40.9%, n = 442) groups, see Table 2.
Patients with unsuccessful cardioversion were more likely to be male (P = .0007), have higher body weight (P = .0006), have coronary artery disease (P = .0365) and receive concomitant use of aspirin (P = .0119) or prior use of statins (defined as statin use between 30 days prior to randomization and the first date of study medication) (P = .0202). Among those on warfarin, INR control was similar between patients with successful or unsuccessful cardioversion, whether assessed as time to achieve therapeutic range or time in therapeutic range (Table 3).
Gender, body weight, body mass index (BMI), baseline CrCl, coronary artery disease history, concomitant use of aspirin, and prior use of statins ( There were no significant differences in primary efficacy regardless of cardioversion success. There were no significant differences in bleeding rates (Table 4) regardless of cardioversion outcome, and, notwithstanding the low numbers, no difference in the edoxaban arm compared with enoxaparin-warfarin.

| S ENS ITIVIT Y ANALYS IS E XCLUDING S P ONTANEOUS C ARDIOVER S ION
For this sensitivity analysis, patients with spontaneous cardioversion were excluded. Patients with unsuccessful cardioversion were more likely to be male (P = .0029), have higher mean weight (P = .0031), have coronary artery disease (P = .0210) and receive concomitant use of aspirin (P = .0075) or prior use of lipid modifying agents (P = .0092, Table 5). There were no significant differences in the primary efficacy outcome regardless of cardioversion success. There were no significant differences in bleeding rates regardless of cardioversion outcome; see Table 6.

| D ISCUSS I ON
In this post hoc analysis from ENSURE-AF, our principal findings are that factors associated with unsuccessful cardioversion were male gender, higher mean weight and higher BMI. Second, efficacy and safety outcomes were low, and were nonsignificantly different whether cardioversion was successful or not.
Predictors associated with successful cardioversion have been investigated, and include various clinical, laboratory, and imaging parameters that have been reported in observational cohorts of varying sizes. 1,4 The strength of this ancillary analysis is evident from its being the largest reported dataset of ECV of AF based on a prospective multicenter trial (ENSURE-AF), with adjudicated outcomes, thus allowing us to determine the factors affecting successful cardioversion.
As expected, approximately one third of the patients had AF recurrence, at a median time of approximately 36-38 days, despite the use of beta blockers (81%) and amiodarone (41%).
In our study, there was no difference in the incidence of amiodarone therapy between the successful cardioversion group and unsuccessful cardioversion group. In one study, 5 amiodarone therapy was considered to be effective in maintaining sinus rhythm following ECV in patient with paroxysmal and/or persistent AF. The efficacy of amiodarone was not influenced by the duration of AF; however, there were only 55 patients enrolled to this study with mean duration of AF > 12 months, thus the study might have been underpowered.
Moreover, data on mean weight and BMI were not available.
The strongest predictor of successful cardioversion in observational cohorts is the duration of prior AF, 4 and in our trial, patients with AF > 12 months in duration were excluded. In addition, elderly age and comorbidities are other clinical factors associated with unsuccessful cardioversion. Age was nonsignificant, but our trial population was relatively young (mean age 64 years).
In our study, male gender was associated with unsuccessful cardioversion. Women were underrepresented in this study.
Possibly men could be affected by additional factors associated with unsuccessful cardioversion, like older age, comorbidities and weight. There is evidence 6 that women are less likely to be treated with cardioversion despite that they tend to be more symptomatic with AF. One of the reasons may be their higher risk of recurrence of AF after successful cardioversion when compared to men. 7,8 Of note, this study clearly shows that body weight predicts unsuccessful cardioversion. This might be because of altered electrical field strengths in this situation. In addition, epicardial fat might also be a factor, which might interfere with the success of ECV. [9][10][11] Moreover, patients with a higher body weight may have higher energy requirements for successful ECV. 6 The concomitant use of aspirin and prior use of statins were also selected as predictors of unsuccessful cardioversion in patients without spontaneous cardioversion, and these are likely to reflect the concomitant comorbidities such as vascular disease.
The above mentioned drugs may also indicate sicker patients.
Indeed, vascular disease is a risk factor for failure of cardioversion within 30 days of ECV 7,12 ; however, the available data are conflicting. In one meta-analysis of randomized controlled trials, atorvastatin was not associated with reduction of the risk of AF recurrence in the secondary prevention subgroup of patients with ECV. 13 Statin use has previously been associated with improved success of cardioversion and maintenance of sinus rhythm. 13,14 In this study, we did not find such an association-but paradoxically, the converse was seen.
Our study shows that higher mean BMI was associated with unsuccessful cardioversion. In the ancillary analysis of ENSURE-AF 15 the BMI < 30 kg/m 2 group had a greater rate of cardioversion success than the BMI ≥ 30 kg/m 2 group. Higher probability of unsuccessful cardioversion in obese patients may reflect associated concomitant diseases or greater body impedance. 16,17 We did not find any significant difference in efficacy and safety outcomes related to the anticoagulation arm, but this is limited by the modest size cohort, limited follow-up, and relatively lower risk. In any case, there is no significant difference between rate and rhythm control strategies, with regard to thromboembolic and bleeding outcomes. 18 We had limited systematic data on echocardiographic parameters, as these were not systematically recorded. Moreover, precise data on the protocol of ECV were not available, as this was at investigator discretion (as ENSURE-AF was an antithrombotic therapy trial, not a trial of a mandated cardioversion protocol). It was generally This was a post hoc ancillary analysis of a clinical trial to assess anticoagulation for ECV in AF, and any outcomes are underpowered and would simply be hypothesis-generating. Nonetheless, this represents data from a contemporary randomized trial with adjudicated clinical outcomes.
To make our post hoc analysis clear, patients were divided into two groups according to maintaining sinus rhythm after cardioversion to evaluate clinical factors related to successful or unsuccessful cardioversion. However, the subgroup with patients without return to sinus rhythm and those who transiently returned to sinus rhythm appear to be heterogeneous.
In conclusion, in this post hoc analysis from ENSURE-AF, our principal findings are that factors associated with unsuccessful cardioversion were male gender, higher mean weight and higher mean BMI. Second, efficacy and safety outcomes were low, and were nonsignificantly different whether cardioversion was successful or Note: The primary efficacy endpoint was the independently adjudicated composite of stroke, SEEs, MI, and CVD occurring between randomization until the end of study in the ITT analysis set.
The primary safety endpoint was the independently adjudicated composite of major and CRNM bleeding, from the time of first administration of study drug to end of treatment + 3 d in the safety analysis set.