Direct oral anticoagulants (DOACs)—including dabigatran, apixaban, rivaroxaban, and edoxaban—are preferred over vitamin K antagonists for the prevention of stroke in patients with nonvalvular atrial fibrillation (AF); unlike vitamin K antagonists, DOACs do not need routine anticoagulation monitoring.1, 2 Edoxaban 60 mg once daily—dose reduced to 30 mg once daily in patients meeting dose-reduction criteria—was approved for the prevention of stroke and systemic embolic events (SEEs) based on the results of the phase 3 Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation Thrombolysis in Myocardial Infarction 48 (ENGAGE AF-TIMI 48) study.3 In this study, edoxaban was noninferior to well-managed warfarin for the prevention of stroke/SEEs and was associated with less major bleeding and cardiovascular (CV)-related death.3 Although warfarin (specifically S-warfarin) is mainly eliminated via CYP2C9-mediated metabolism with minimal contribution of renal clearance, edoxaban, like all other DOACs, is partially eliminated via the kidney, with ≈50% renal clearance (dabigatran 80% renally cleared, rivaroxaban 35%, and apixaban 27%).4, 5 Given the high renal clearance, it was postulated that edoxaban exposure may be lower in patients with high creatinine clearance (CrCL >100 mL/min); thus, efficacy could hypothetically be lower in this subpopulation.

The impact of renal function on response to edoxaban was explored in subgroup analyses by Bohula et al of the ENGAGE AF-TIMI 48 study (ENGAGE AF).6 It was noted that in patients with AF with CrCL > 95 mL/min, there was a trend toward decreased relative efficacy of edoxaban 60 mg once daily compared with warfarin (stroke/SEE hazard ratio 1.36; 95%CI, 0.88-2.10; P = 0.17), although the absolute event rate in the edoxaban 60-mg subgroup was low (1.1%/year).6 However, it is important to note that in the ENGAGE AF study, patients in the warfarin treatment group with high CrCL had substantially lower ischemic stroke event rates than, and not consistent with, the published warfarin ischemic stroke event rates in this subpopulation in the other 3 major DOAC AF pivotal trials.3, 7-10 Thus, the observed unfavorable hazard ratio for the stroke event rate in ENGAGE AF in patients treated with edoxaban 60 mg once daily relative to warfarin in this limited select subgroup of patients with CrCL >95 mL/min appears driven by a surprisingly low event rate for the warfarin arm.3

To further characterize the impact of high CrCL on edoxaban exposure, this prospective study was conducted to compare the drug exposure of edoxaban 60 mg once daily vs edoxaban 75 mg once daily in patients with AF who were anticoagulant-naïve and had a CrCL >100 mL/min.

Methods Study Design

This was a prospective, randomized, double-blind study with evaluation of end points by an independent clinical event committee. Patients were randomized (1:1) via an interactive web/voice response system to receive either edoxaban 75 mg once daily or edoxaban 60 mg once daily. Patients in both arms were treated for up to 12 months with a 2- to 4-week follow-up period. Treatment interruptions were discouraged but could occur due to an adverse event (AE) or other medical reasons. No edoxaban dose reductions were allowed in the study, and no patients who met standard labeling criteria for edoxaban dose reduction were enrolled.

The on-treatment period was defined as the time during which patients were taking the study drug through up to 3 days after their last dose. The overall study period was the time from the date of the first dose of the study drug to the follow-up visit. For patients without a follow-up visit, the overall study period was defined as the final dose plus 28 days. A summary of the study design is presented in Figure 1.

Study design. The goal was to randomize ≈600 patients (300 per treatment arm—edoxaban 60 mg once daily or edoxaban 75 mg once daily) with a treatment duration of 12 months. CHADS2 indicates stroke risk stratification scheme for patients with atrial fibrillation (congestive heart failure, hypertension, age, diabetes, and previous stroke); CrCL, creatinine clearance; NVAF, nonvalvular atrial fibrillation; R, randomization.

The protocol and all study documents were approved by the appropriate institutional review boards and independent ethics committees. The full list of study sites, the majority of which were in eastern Europe, and their corresponding approval committee is in Table S1. This study was conducted in accordance with the Good Clinical Practice standards for drugs and the ethical principles specified in the Declaration of Helsinki. Informed consent was obtained from all patients before enrollment.

Patient Population

Patients were required to have a diagnosis of nonvalvular AF documented by any electrical tracing within the prior 12 months and a CrCL >100 mL/min (calculated by the Cockcroft-Gault formula),11 and patients were required to be anticoagulant-naïve (defined as having received no dose of any oral anticoagulants for 30 days before randomization) with a stroke risk stratification scheme for patients with atrial fibrillation (CHADS2) score of ≥2. Patients with an indication for an edoxaban dose reduction (body weight <60 kg, CrCL <50 mL/min, or the use of concomitant P-glycoprotein inhibitors) were excluded. Additional exclusion criteria included concomitant medication use such as nonstudy anticoagulants, or chronic oral, or parenteral nonaspirin/nonsteroidal anti-inflammatory drugs for ≥4 days/week. A complete list of exclusion criteria is summarized in Table S2.

Study Objectives and Assessments

The primary pharmacokinetic (PK) parameters were the minimum and average edoxaban plasma concentrations at steady state; anti-factor Xa (FXa) activity was the primary pharmacodynamic (PD) parameter. PK/PD samples were collected before dosing, 1 to 2 hours after dosing, and 4 to 8 hours after dosing on days 30, 90, and 360 in the lithium heparin tube (for PK) and the sodium citrate tubes (for PD). Edoxaban plasma concentration was analyzed using a validated liquid chromatographic–tandem mass spectrometric method in Q2 Solutions (Ithaca, New York).12 Anti-FXa levels were measured at Medpace Research Laboratory (Cincinnati, Ohio) using a commercially available anti-FXa activity assay (STA-Liquid Anti-Xa; Diagnostica Stago, Asnières sur Seine, France) with an edoxaban-specific setup using the STA-Edoxaban Calibrator and STA-Edoxaban Control on the STA-R analyzer (Diagnostica Stago).13

To account for variability in PK parameters due to sample collection within time windows rather than at precise time points in the treatment arms, edoxaban steady-state PK values were estimated by applying a previously developed population PK (popPK) model for edoxaban that used pooled data from 13 phase 1 studies along with the phase 3 study.14 This popPK model included body weight, race (Asian vs non-Asian), health status (healthy patients vs patients with nonvalvular AF), CrCL (mL/min), and P-glycoprotein inhibitor as covariates on relevant PK parameters. Model-estimated steady-state concentrations were used to make precise and adequate comparisons of PK exposure parameters between the 2 doses.

Clinical efficacy and safety outcomes events were tabulated and reported. Efficacy outcomes included the composite of stroke, transient ischemic attack (TIA), and SEEs; the composite of ischemic stroke, TIA, and SEEs; and the composite of stroke and/or TIA, SEEs, myocardial infarction, CV death, and major bleeding. Secondary safety outcomes included the incidence of major (including intracranial) and clinically relevant nonmajor (CRNM) bleeding, and all bleeding events that were categorized as major, CRNM, or nuisance bleeding; AEs of special interest included combined elevations of aminotransferases and bilirubin (alanine or aspartate aminotransferase ≥3 times the upper limit of normal and total bilirubin ≥2 times the upper limit of normal). All investigator-reported clinical outcome events were adjudicated by an independent panel of experts without knowledge of the study treatment. The adjudicated bleeding events were defined as major bleeding, CRNM bleeding, nuisance bleeding, or no bleeding event, based on International Society on Thrombosis and Haemostasis classification.

Statistical Analysis

Continuous data were summarized with descriptive statistics. Categorical data were summarized with frequency counts and percentages. The percentages were calculated based on the total number of patients in each treatment group and analysis category.

The safety analysis set included all patients who were randomized and who received at least 1 dose of the study drug. The PK analysis set (PKAS) included all patients in the safety analysis set who received edoxaban and had at least 1 postdose PK concentration measurement. The PD analysis set included all patients in the safety analysis set who had at least 1 predose PD assessment and at least 1 postdose PD assessment. The modified intent-to-treat (mITT) analysis set included all randomized patients who received at least 1 dose of the study drug.

Results Patient Disposition and Demographics

Overall, 607 patients were randomized to edoxaban 60 mg once daily or edoxaban 75 mg once daily (303 in the edoxaban 60 mg once-daily arm and 304 in the edoxaban 75 mg once-daily arm). The majority of randomized patients (98.2% [596/607]) completed the study, with similar proportions of completing patients in both treatment groups. Eleven patients withdrew from the study (10 due to death; 1 patient withdrew from the study after 173 days). No patients were lost to follow-up.

The demographics and baseline clinical characteristics were well balanced between treatment arms (Table 1). The demographics and baseline characteristics were predominately younger (median age, 61), more likely men (72%), with higher body mass index (median, 34.6) and more extensive medical history of congestive heart failure, hypertension, and diabetes mellitus (Table 1) than the overall AF population in ENGAGE AF-TIMI 48.3 Of note, the demographics and baseline characteristics of patients in this study were similar to the patients with high CrCL enrolled in ENGAGE AF-TIMI 48 (Table S4). Five patients had an initial CrCL > 100 mL/min via local lab measurement at randomization but were subsequently found to have CrCL < 100 mL/min via central lab measurement (Table 1); the CrCL for these patients ranged from 92 to 100 mL/min. Of these 5 patients, 1 received edoxaban 60 mg once daily, and 4 received edoxaban 75 mg once daily. None experienced a stroke, bleeding event, or serious AE during the study.

Table 1. Demographic and Baseline Clinical Characteristics (mITT Analysis Set) Edoxaban 60 mg Once Daily (n = 303) Edoxaban 75 mg Once Daily (n = 303) Overall (N = 606) Age, y Mean (SD) 60.9 (8.1) 60.4 (8.4) 60.6 (8.2) Median (min-max) 61.0 (32-80) 61.0 (32-82) 61.0 (32-82) <65, n (%) 195 (64.4) 201 (66.3) 396 (65.3) 65 to <75, n (%) 97 (32.0) 91 (30.0) 188 (31.0) ≥75, n (%) 11 (3.6) 11 (3.6) 22 (3.6) Male, n (%) 215 (71.0) 221 (72.9) 436 (71.9) BMI, kg/m2 Mean (SD) 35.4 (5.8) 35.3 (6.4) 35.4 (6.1) Median (min-max) 35.1 (23.2-55.2) 34.4 (22.0-58.3) 34.6 (22.0-58.3) <30, n (%) 47 (15.5) 55 (18.2) 102 (16.8) ≥30 to <35, n (%) 102 (33.7) 113 (37.3) 215 (35.5) ≥35, n (%) 154 (50.8) 133 (43.9) 287 (47.4) CrCL, mL/min Mean (SD) 128.7 (24.2) 129.4 (28.0) 129.0 (26.2) Median (min-max) 123.4 (100.0-250.8) 122.3 (92.0-295.8) 122.5 (92.0-295.8) ≤100, n (%) 1 (0.3) 4 (1.3) 5 (0.8) >100 to ≤120, n (%) 134 (44.2) 137 (45.2) 271 (44.7) >120, n (%) 168 (55.4) 162 (53.5) 330 (54.5) Atrial fibrillation type, n (%) Paroxysmal 122 (40.3) 134 (44.2) 256 (42.2) Persistent 87 (28.7) 82 (27.1) 169 (27.9) Permanent 94 (31.0) 87 (28.7) 181 (29.9) CHADS2 Mean (SD) 2.5 (0.8) 2.5 (0.8) 2.5 (0.8) Median (min-max) 2.0 (2.0-5.0) 2.0 (1.0-5.0) 2.0 (1.0-5.0) <2, n (%) 0 1 (0.3) 1 (0.2) 2, 3, n (%) 268 (88.4) 270 (89.1) 538 (88.8) >3, n (%) 35 (11.6) 32 (10.6) 67 (11.1) Medical history, n (%) Congestive heart failure 231 (76.2) 231 (76.2) 462 (76.2) Hypertension 300 (99.0) 299 (98.7) 599 (98.8) Diabetes 112 (37.0) 130 (42.9) 242 (39.9) Ischemic/embolic stroke 30 (9.9) 24 (7.9) 54 (8.9) BMI, body mass index; CHADS2, stroke risk stratification scheme for patients with atrial fibrillation (congestive heart failure, hypertension, age, diabetes, previous stroke); CrCL, creatinine clearance; mITT, modified intent-to-treat; SD, standard deviation. Pharmacokinetic End Points

The PKAS included 298 patients randomly assigned to edoxaban 60 mg once daily and 297 patients randomly assigned to edoxaban 75 mg once daily. Figure S1 shows the observed edoxaban plasma concentrations before dosing, 1 to 2 hours after dosing, and 4 to 8 hours after dosing on days 30, 90, and 360 of the on-treatment periods for both treatment arms. As expected, the resulting measured actual plasma concentrations showed large variability, partly due to noise from different sampling times within the specified collection window. The popPK model adequately described the observed steady-state concentrations during the on-treatment period among patients included in the PKAS (Figure S2). The model-estimated steady-state edoxaban exposure was ≈25% higher in patients receiving 75 mg once daily, as compared with patients receiving 60 mg once daily (Table 2). The average edoxaban concentration at steady state was 74.8 ng/mL in the 60-mg group and 93.2 ng/mL in the 75-mg group (Table 2).

Table 2. Population Pharmacokinetic Parameters Estimates for Edoxaban at Steady State by Dose (Pharmacokinetic Analysis Set) Edoxaban 60 mg Once Daily (n = 298) Edoxaban 75 mg Once Daily (n = 297) Cav (ng/mL) Mean (SD) 74.8 (13.3) 93.2 (16.2) Median (min-max) 74.5 (44.5-121) 92.4 (55.5-146) Geometric mean/CV (%) 73.7/17.8 91.8/17.8 Ratio of means edoxaban 75/60 mg (95%CI) 1.25 (1.2-1.3) P value <.0001 Cmax, ng/mL Mean (SD) 214 (51.1) 269 (62.4) Median (min-max) 211 (124-429) 261 (144-496) Geometric mean/CV (%) 209/23.7 262/23.2 Ratio of means edoxaban 75/60 mg (95%CI) 1.25 (1.2-1.3) P value <.0001 Cmin, ng/mL Mean (SD) 18.2 (5.2) 22.5 (6.5) Median (min-max) 17.8 (4.8-33.9) 22.1 (8.7-42.4) Geometric mean/CV (%) 17.4/30.6 21.6/30.5 Ratio of means edoxaban 75/60 mg (95%CI) 1.24 (1.2-1.3) P value <.0001 Cav, average plasma concentration of edoxaban at steady state; Cmax, maximum (peak) concentration of drug in blood plasma (applied to extravascular drug administration); Cmin, minimum observed (or could infer lowest effective) concentration of edoxaban in blood plasma; CV, coefficient of variation; min, minimum; max, maximum; SD, standard deviation. Anti-FXa

The mean predose concentrations of anti-FXa in patients in the PD analysis set during days 30, 90, and 360 of the on-treatment period ranged from 23.1 to 28.0 ng/mL in the edoxaban 60-mg once-daily arm and 33.4 to 35.9 ng/mL in the edoxaban 75-mg once-daily arm. Anti-FXa levels for edoxaban 60 mg once daily and 75 mg once daily at days 30, 90, and 360 are provided in Table 3. Figure 2 illustrates the linear correlation between anti-FXa activity and edoxaban concentration.

Table 3. Anti-FXa (ng/mL) Measures at Days 30, 60, and 360 in Patients Receiving Edoxaban 60 mg Once Daily or Edoxaban 75 mg Once Daily Edoxaban 60 mg Once Daily (n = 292) Edoxaban 75 mg Once Daily (n = 292) Before Dosing 1-2 h After Dosing 4-8 h After Dosing Before Dosing 1-2 h After Dosing 4-8 h After Dosing Day 30 N 284 280 280 284 291 292 Mean (SD) 28 (58) 201 (109) 161 (63) 33 (58) 233 (122) 191 (73) Median (min-max) 0 (0-400) 205 (0-400) 145 (27-400) 22 (0-397) 251 (0-400) 181 (0-400) Day 90 N 281 281 283 273 280 277 Mean (SD) 23 (47) 180 (107) 157 (67) 33 (59) 229 (120) 184 (79) Median (min-max) 0 (0-375) 173 (0-400) 144 (0-363) 22 (0-386) 234 (0-400) 174 (0-400) Day 360 N 276 277 279 269 269 267 Mean (SD) 26 (55) 181 (111) 147 (77) 356 (59) 215 (126) 184 (95) Median (min-max) 0 (0-399) 166 (0-400) 139 (0-400) 23 (0-387) 215 (0-400) 177 (0-400) FXa, factor Xa; min-minimum; max, maximum; SD, standard deviation.

Scatterplot of edoxaban concentration versus anti-FXa in the PD analysis set. Anti-FXa values greater than the upper limit of quantification (400 ng/mL) were reported as 400 ng/mL, and anti-FXa values less than the lower limit of quantification (20 ng/mL) were reported as 0 ng/mL. FXa indicates factor Xa; PD, pharmacodynamics.

A post hoc analysis was also performed to exclude the patients who had measurable anti-FXa at baseline (likely due to concomitant heparin use). With this exclusion, the change in anti-FXa levels also remained consistent with the increase in dose.

Adjudicated Efficacy Outcomes

In the mITT population during the on-treatment period, the incidence of the composite end point of stroke, TIA, and SEEs was 0.7% (2/303) in the edoxaban 60-mg once-daily arm vs 1.0% (3/303) in the edoxaban 75-mg arm (odds ratio, 1.51; 95%CI, 0.2-18.1; P = 1.000; Table 4). All events were stroke events; there were no reported TIA or SEE events. In the edoxaban 60-mg once-daily group, two patients experienced an ischemic stroke; in the 75-mg once-daily group, 1 patient experienced an ischemic stroke and 2 experienced a hemorrhagic stroke (1 fatal). No other patients in the mITT population experienced an event during the overall study period. The annual event rate of the composite end point was <1% during the study.

Table 4. Adjudicated Clinical Outcome Events (on Treatment) by Dose Edoxaban 60 mg Once Daily (n = 303) Edoxaban 75 mg Once Daily (n = 303) Overall (N = 606) Stroke, TIA, or SEE, n (%) – mITT 2 (0.7) 3 (1.0) 5 (0.8) Ischemic stroke 2 1 3 Hemorrhagic stroke 0 2a 2 TIA or SEE 0 0 0 Patient exposure years 307 303 610 Annualized event rate, % 0.7 1.0 0.9 95% exact CI for annualized event rate, % (0.1-2.4) (0.2-2.9) (0.3-1.9) Odds ratio (95% exact CI): 75 mg vs 60 mg

1.51 (0.17-18.1)

1.00

P value Major or CRNM bleeding, n (%) – SAS 11 (3.6) 10 (3.3) 21 (3.5) Patient exposure years 292 287 579 Annualized event rate, % 3.8 3.5 3.6 95% exact CI for annualized event rate, % (1.9-6.8) (1.7-6.4) (2.2-5.5) Odds ratio (95% exact CI): 75 mg vs 60 mg

0.91 (0.34-2.39)

1.00

P value Major bleeding, n (%) – SAS 2 (0.7) 3 (1.0) 5 (0.8) Intracranial 0 2 (0.7)a 2 (0.3) Upper gastrointestinal 0 1 (0.3) 1 (0.2) Patient exposure years 293 289 582 Annualized event rate, % 0.7 1.0 0.9 95% exact CI for annualized event rate, % (0.1-2.5) (0.2-3.0) (0.3-2.0) Odds ratio (95% exact CI): 75 mg vs 60 mg

1.51 (0.17-18.1)

1.00

P value Net clinical end point,b n (%) – mITT 5 (1.7) 8 (2.6) 13 (2.1) Patient exposure years 293 288 581 Annualized event rate, % 1.7 2.8 2.2 95% exact CI for annualized event rate, % (0.6-4.0) (1.2-5.5) (1.2-3.8)