Clinical Trials

The systematic review included one RCT of patients with nonvalvular atrial fibrillation. ENGAGE AF-TIMI 48 (N = 21,105) evaluated the non-inferiority (NI) of edoxaban to warfarin in patients with atrial fibrillation. Patients in the ENGAGE AF-TIMI 48 study were randomized to warfarin, edoxaban 60 mg once daily, or edoxaban 30 mg once daily. Patients randomized to the warfarin group received dose-adjusted warfarin to maintain a monthly measured international normalized ratio (INR) at 2.0 to 3.0. Patients randomized to edoxaban 60 mg or 30 mg groups received the allocated dose unless one of the following criteria was present, in which case the edoxaban dose was halved: creatinine clearance of 30 mL to 50 mL per minute, body weight of 60 kg or less, or verapamil or quinidine use. All patients received two medications — the active allocated drug, and a matched placebo for the non-allocated drug. To maintain blinding, sham INR values were generated to patients who were randomized to edoxaban. The intended duration depended on collecting a total of 672 events of the primary outcome; the end-result median duration of treatment was 2.5 years. Upon the completion of the study, patients went through a transition period to an anticoagulant chosen by the treating physician. The primary efficacy outcome was a composite of stroke and SEE (time to first adjudicated ischemic or hemorrhagic stroke, or SEE).

NI would be demonstrated if the upper boundary of the one-sided 97.5% CI of edoxaban compared with warfarin was below the NI margin of 1.38 for the HR. If NI was achieved, the primary outcome of stroke/SEE was then tested at an alpha of 0.01 for superiority; if superiority was achieved, the secondary outcome of stroke/SEE/CV mortality was tested for superiority at an alpha of 0.01; if superiority was achieved, the secondary outcome of major adverse cardiovascular events, which includes myocardial infarction, stroke, SEE, or CV mortality was tested for superiority at an alpha of 0.01; and if superiority was achieved, the secondary outcome of stroke/SEE/all-cause mortality was tested for superiority.

Outcomes

Outcomes were defined a priori in the CDR systematic review protocol. Of these, CDEC discussed the following:

• composite of stroke and SEE
• ischemic stroke
• hemorrhagic stroke
• systemic thromboembolic event
• major bleeding.

Efficacy

• Edoxaban was shown to be noninferior, but not superior, to warfarin for the primary efficacy outcome of stroke and SEE. The use of 60 mg edoxaban was associated with the following HRs compared with warfarin:
  • Modified, on-treatment period, intention-to-treat population: 0.79 (97.5% CI, 0.63 to 0.99).
  • Overall study period, intention-to-treat population: 0.87 (97.5% CI, 0.73 to 1.04).
• During the treatment period, the proportion of patients (modified intention to treat) who experienced individual events classified as stroke were as follows (60 mg edoxaban versus warfarin):
  • ischemic stroke: 0.87% versus 0.93%
  • hemorrhagic stroke: 0.26% versus 0.49%
  • fatal stroke: 0.29% versus 0.28%
  • disabling stroke: 0.23% versus 0.26%
  • SEE: 0.05% versus 0.08%.

Harms (Safety and Tolerability)

• The safety outcome, major bleeding, was presented in the ENGAGE AF-TIMI 48 study as a pre-specified primary safety outcome. The incidence of major bleeding (modified International Society on Thrombosis and Haemostasis criteria) was lower with edoxaban 60 mg versus warfarin in (418 patients, 2.75% versus 524 patients, 3.43%; HR = 0.80 [95% CI, 0.71 to 0.91]). The incidences of major bleeding designated as fatal bleeding or hemorrhagic stroke were lower with edoxaban 60 mg versus warfarin, while there were more patients with gastrointestinal hemorrhage in the edoxaban 60 mg group than in the warfarin group.

The incidence of non-bleeding adverse events was similar between edoxaban 60 mg and warfarin groups. Approximately one-third of the patients in the ENGAGE AF-TIMI 48 trial experienced a non-bleeding serious adverse event (SAE). Numerically, 2,315 patients (33.0%) experienced non-bleeding SAEs in the edoxaban 60 mg group, and 2,516 patients (35.9%) experience non-bleeding SAEs in the warfarin group. The number of patients who discontinued the allocated treatment due to adverse events was similar in the edoxaban 60 mg group and in the warfarin group. Overall, there were fewer deaths in the edoxaban 60 mg group (769 patients, 11.0%) than in the warfarin group (836 patients, 11.9%). In both groups, the bulk of the deaths were driven by cardiovascular causes (7.5% in the edoxaban 60 mg groups, and 8.7% in the warfarin group).

Cost and Cost-Effectiveness

Edoxaban (Lixiana) is available as a 30 mg and 60 mg tablet at a submitted price of $2.84, regardless of dosage strength. At a recommended dose of 60 mg once daily, with the potential for dose reduction to 30 mg once daily if required, the daily cost of treatment is $2.84.

The manufacturer submitted a cost-utility analysis based on a Markov model, compared with a lifetime horizon, from the perspective of the publicly funded health care system. Analyses were conducted comparing edoxaban with warfarin (5 mg once daily) and rivaroxaban (60 mg once daily). For the comparison with warfarin, data from the ENGAGE AF-TIMI 48 trial were used; while for the comparison with rivaroxaban, data from a manufacturer-submitted network meta-analysis (NMA) were used. Analyses comparing edoxaban with dabigatran (110 mg twice daily and 150 mg twice daily) and apixaban (5 mg twice daily) are included as sensitivity analyses, with the necessary clinical data obtained from the manufacturer-submitted NMA. The rationale for excluding these comparators from the primary analysis was two-fold: limited comparability of the ENGAGE AF-TIMI 48 trial of edoxaban and the pivotal trials for dabigatran and apixaban; as well as rivaroxaban being the most commonly used DOAC in Canada. According to the manufacturer’s base-case analysis, treatment with edoxaban was more effective (quality-adjusted life-year [QALY] gain of 0.1517) and more costly ($1,922) than warfarin leading to an incremental cost per QALY gained of $12,672. Edoxaban was dominant compared with rivaroxaban (i.e., it was less costly and more effective).

No major limitations with respect to the model, assumptions, and data inputs were found. Limitations relate to the context of the decision problem:

• Given the patient population from the ENGAGE AF-TIMI trial, analysis is only relevant to AF patients with CHADS₂ ≥ 2; whereas edoxaban may be used in patients with CHADS₂ = 1.
• The submitted analysis highlights only the comparisons of edoxaban with warfarin and rivaroxaban. Given current funding of DOACs, comparisons with apixaban and dabigatran are also relevant and should have been given equal prominence.

CDR reanalysis incorporating all relevant comparators (warfarin, other DOACs) found, for patients with nonvalvular atrial fibrillation (CHADS₂ ≥ 2) requiring anticoagulation, apixaban was the most cost-effective DOAC and all other new oral anticoagulant (NOACs) (including edoxaban) were dominated (i.e., less effective and more costly). Apixaban remained cost-effective compared with edoxaban (at a willingness-to-pay threshold of $50,000) unless the price of edoxaban was reduced by at least 33%. It was noted that the relative cost-effectiveness of edoxaban versus apixaban and dabigatran is somewhat uncertain due to limitations of the clinical data.