Cardioversion of Atrial Fibrillation in ENGAGE AF‐TIMI 48

Introduction

Three post‐hoc analyses from warfarin‐controlled trials of non–vitamin K oral anticoagulants in patients with atrial fibrillation (AF)1 and 1 prospective study2 comparing rivaroxaban and warfarin in 1504 patients have shown low and comparable rates of thromboembolic events and hemorrhagic complications in patients undergoing electrical cardioversion. We compared outcomes with edoxaban vs warfarin in the Effective Anticoagulation With Factor Xa Next Generation (ENGAGE) AF‐TIMI 48 trial among patients who underwent a first electrical cardioversion while on study drug.3

Methods

This 3‐arm, randomized, double‐blind, double‐dummy, global phase 3 clinical trial compared a once‐daily higher‐dose edoxaban (60 mg/30 mg) regimen, a lower‐dose edoxaban (30 mg/15 mg) regimen, and warfarin (target international normalized ratio [INR] of 2.0–3.0) in 21 105 patients with AF and CHADS₂ (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, prior stroke/transient ischemic attack/thromboembolism) score ≥2. Patients with moderate renal dysfunction, weight ≤60 kg, or concomitant use of P‐glycoprotein inhibitors received a 50% dose reduction of edoxaban (or matching edoxaban‐placebo). Major exclusion criteria were AF due to reversible cause, severe renal impairment, increased bleeding risk, mechanical heart valve, or moderate to severe mitral stenosis. Both edoxaban regimens were noninferior to warfarin in preventing the composite of stroke (including both hemorrhagic and nonhemorrhagic stroke) or systemic embolic events (SEE) while significantly reducing major bleeding, intracerebral hemorrhage, and cardiovascular mortality.

Results

There were 832 attempted electrical cardioversions performed to restore sinus rhythm in patients with AF or atrial flutter (we considered all shocks delivered on the same date as 1 cardioversion attempt) during the trial. We excluded 200 cardioversions that occurred >3 days after the most recent dose of blinded study anticoagulant (edoxaban or warfarin; ie, during a period of interruption of blinded study drug of ≥4 days). There were a total of 632 electrical cardioversion attempts performed while on study drug in 365 patients occurring on average 348 days (interquartile range [IQR], 86–526 days) after randomization. Compared with others in the trial, the 365 patients undergoing a first cardioversion while on study drug were, at the time of randomization, significantly younger (median age, 70 vs 72 years) and more likely to have paroxysmal AF (47% vs 25%), a CHADS₂ score ≤3 (88% vs 77%), and to have been enrolled in North America (41% vs 22%; P < 0.001 for each). There were no differences between randomized treatment groups in baseline characteristics among these 365 patients. In the warfarin group, the median time between the last INR measurement and cardioversion was 6 days (IQR, 1–13 days) and median INR value was 2.5 (IQR, 2.1–3.0).

In the 30 days after cardioversion, stroke or SEE occurred in 2 patients on the lower‐dose edoxaban regimen; none occurred with warfarin or higher‐dose edoxaban (Table 1). There were no major bleeding events and 1 death (higher‐dose edoxaban) in the 30 days after cardioversion. From cardioversion to the end of the trial (median follow‐up 25 months), there were no statistically significant differences in primary efficacy and safety endpoints between treatment groups. No additional events occurred in patients who had multiple cardioversions while on study drug (267 additional cardioversions) in the 30 days post–repeat cardioversions.

Table 1.

Clinical Events in the First 30 Days After First Cardioversion While on Treatment

	Warfarin, n = 114	HDE, n = 140	LDE, n = 111
Stroke or SEE, n (%)	0	0	2 (1.81)
Major bleeding, n	0	0	0
Death, n (%)	0	1 (0.71)	0

Abbreviations: HDE, higher‐dose edoxaban; LDE, lower‐dose edoxaban; SEE, systemic embolic events.

Discussion

In this post‐hoc analysis of the ENGAGE AF‐TIMI 48 trial, thromboembolic and major bleeding events in the 30 days post‐cardioversion were infrequent (<2%) and similar in the treatment and control groups. Likewise, no differences between treatment groups were seen post‐cardioversion throughout the end of the trial. These low event rates are similar to the rates (≤2%) reported in the other warfarin‐controlled trials of non–vitamin K oral anticoagulants.1, 2

Study Limitations

Our results are descriptive, and although comparisons are made, they should be interpreted with caution. This was a post‐hoc, nonrandomized comparison between edoxaban and warfarin; patients undergoing cardioversion in the trial were at lower risk than the overall trial population; and data from transesophageal echocardiography were not collected during the trial. The low event rates observed and limited power in this analysis indicate that a much larger cohort would be needed to achieve adequate study power to evaluate differences between oral anticoagulants.

Future Directions

A prospective, randomized, open‐label, parallel‐group, phase 3b trial4—the Edoxaban vs. Warfarin in Subjects Undergoing Cardioversion of Atrial Fibrillation (ENSURE‐AF, NCT02072434)—is being conducted in >2000 patients to prospectively study the efficacy and safety of edoxaban in patients undergoing cardioversion. The primary objectives of the trial are to compare rates of stroke, SEE, myocardial infarction, and cardiovascular mortality in subjects receiving edoxaban with those receiving enoxaparin/warfarin.

Conclusion

Thromboembolic and major bleeding events post cardioversion were infrequent and similar with edoxaban and warfarin in the ENGAGE AF‐TIMI 48 trial.