Edoxaban vs warfarin in patients with nonvalvular atrial fibrillation in the US Food and Drug Administration approval population: An analysis from the Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation–Thrombolysis in Myocardial Infarction 48 (ENGAGE AF–TIMI 48) trial

Edoxaban vs warfarin in patients with nonvalvular atrial fibrillation in the US Food and Drug Administration approval population: An analysis from the Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation–Thrombolysis in Myocardial Infarction 48 (ENGAGE AF–TIMI 48) trial Alon Eisen, MD, a Robert P. Giugliano, MD, SM, a Christian T. Ruff, MD, MPH, a Francesco Nordio, PhD, a Harinder S. Gogia, MD, b Vivek R. Awasty, MD, c David A. Henderson, MD, d Michele F. Mercuri, MD, PhD, e Howard Rutman, MD, f Elliott M. Antman, MD, a and Eugene Braunwald, MD a Boston, MA; Anaheim, CA; Marion, OH; Daytona Beach, FL; Edison, and Parsippany, NJ

Background

Edoxaban is a specific anti-Xa inhibitor that, in comparison to warfarin, has been found to be noninferior for the prevention of stroke or systemic embolism (SSE) and to reduce bleeding significantly in patients with nonvalvular atrial fibrillation (AF). The US Food and Drug Administration (FDA) approved the higher-dose edoxaban regimen (60/30 mg) in patients with AF and a creatinine clearance of ≤95 mL/min. We report for the first time the clinical characteristics, efficacy, and safety of the FDA-approved population in the ENGAGE AF–-TIMI 48 trial.

Methods The patients included had been treated with either warfarin or edoxaban 60/30 mg and had a creatinine clearance of ≤95 mL/min. The primary efficacy was SSE, and the principal safety end point was major bleeding (International Society on Thrombosis and Haemostasis classification). Median follow-up was 2.8 years. Results Patients in the FDA-approved cohort were older, were more likely female, and had higher CHADS2 and HAS-BLED scores, as compared with patients not included in the FDA label. The primary end point occurred in 1.63%/y with edoxaban vs 2.02%/y with warfarin (hazard ratio [HR] 0.81, 95% CI 0.67-0.97, P = .023). Edoxaban significantly reduced the rate of hemorrhagic stroke (HR 0.47, 95% CI 0.31-0.72, P b .001) and cardiovascular death (HR 0.84, 95% CI 0.73-0.97, P = .015). Ischemic stroke rates were similar between the treatment groups (1.31%/y vs 1.39%/y, P = .97). Major bleeding was significantly lower with edoxaban (3.16%/y vs 3.77%/y; HR 0.84, 95% CI 0.72-0.98, P = .023). Conclusion In the FDA-approved cohort of the ENGAGE AF–-TIMI 48 trial, treatment with edoxaban 60/30 mg was superior to warfarin in the prevention of SSE and significantly reduced cardiovascular death and bleeding, especially fatal bleeding and hemorrhagic stroke. (Am Heart J 2016;172:144-151.)

From the aTIMI Study Group, Division of Cardiovascular Medicine, Brigham and Women's b

Hospital, and Department of Medicine, Harvard Medical School, Boston, MA, Anaheim Regional Medical Center, Anaheim, CA, cMarion General Hospital, Marion, OH, dFlorida Hospital Memorial Medical Center, Daytona Beach, FL, eDaiichi Sankyo Pharma Development, Edison, NJ, and fDaiichi Sankyo Inc, Parsippany, NJ. RCT #NCT00781391. Submitted August 26, 2015; accepted November 12, 2015. Reprint requests: Eugene Braunwald, MD, TIMI Study Group, 350 Longwood Ave, 1st Floor Offices, Boston, MA 02115. E-mail: [email protected] 0002-8703 © 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ahj.2015.11.004

Edoxaban is an oral, reversible, direct, and specific factor Xa inhibitor. In the ENGAGE AF–TIMI 48 trial, 1 both once-daily dose regimens of edoxaban (60/30 and 30/15 mg) were noninferior but not superior to wellmanaged warfarin (median time in therapeutic range [TTR] 68.4%) in reducing the risk of stroke (ischemic or hemorrhagic) or systemic embolic event (SSE) in patients with nonvalvular atrial fibrillation (AF). Both edoxaban regimens significantly reduced major and minor bleeding events compared with warfarin, including fatal and intracranial hemorrhage.

American Heart Journal Volume 172, Number 0

In the ENGAGE AF–TIMI 48 trial, a positive correlation was demonstrated between edoxaban concentration, anti-Xa activity, and clinical efficacy. 2 In healthy volunteers, about 50% of absorbed edoxaban is eliminated by renal clearance. 3 As might be expected, pharmacokinetic studies demonstrated that the systemic exposure to edoxaban is inversely related to renal function; as creatinine clearance (CrCl) increases, edoxaban exposure decreases, 4,5 which raises the possibility that its efficacy could become attenuated. In a prospective pharmacokinetic substudy of the ENGAGE AF–TIMI 48 trial, edoxaban trough concentrations on day 29 were about 40% lower in patients with CrCl N 95 mL/min when compared with patients with a CrCl of N50 to ≤80 mL/min. 6 Given these findings, the US Food and Drug Administration (FDA) approved the higher-dose regimen given once daily for patients with AF at moderate to high risk for stroke and with CrCl 15-95 mL/min (60 mg if CrCl 50-95 mL/min; 30 mg if CrCl 15-50 mL/min). 6 In addition, the FDA label for AF did not recommend dose reduction for patients of low body weight or concomitant use of potent P-glycoprotein inhibitors, as was specified in the protocol of the ENGAGE AF–TIMI 48 trial. 1 We present here, for the first time, the clinical efficacy and safety of edoxaban 60/30 mg in the cohort of AF patients in ENGAGE AF–TIMI 48 treated as per the FDA-approved drug label.

Methods The ENGAGE AF–TIMI 48 study was a 3-group, randomized, double-blind, double-dummy multinational trial that compared 2 dosing regimens of edoxaban with warfarin in 21,105 patients with AF1 (NCT00781391). A detailed description of the study population, randomization, and study design had been described previously. 7 In brief, eligible patients were 21 years or older who had AF within 12 months prior to randomization, a CHADS2 score of 2 or above, and planned treatment with anticoagulation. Key exclusion criteria included AF due to a reversible cause, mechanical valves or moderate-severe mitral stenosis, increased risk of bleeding, recent acute coronary syndrome or stroke within 30 days prior to randomization, and the use of dual antiplatelet therapy. Patients were randomly assigned to receive warfarin, with continual dose adjustment to achieve an international normalized ratio of 2.0 to 3.0; edoxaban 60 mg once daily; or edoxaban 30 mg once daily. For patients in either edoxaban group, the dose was halved if one or more of the following characteristics associated with anticipated increased drug exposure were present during randomization or during the study: calculated CrCl 30-50 mL/min, body weight ≤60 kg, or the concomitant use of the potent P-glycoprotein inhibitors verapamil, quinidine, and dronedarone. The primary efficacy end point was defined as SSE. The principal safety end point was major bleeding, as defined by the International Society on Thrombosis and Haemostasis classification. 8 The primary net clinical outcome was

Eisen et al 145

a composite of SSE, major bleeding, or death from any cause. Additional net clinical outcomes were a composite of disabling stroke, life-threatening bleeding, or death from any cause, and an exploratory composite of SSE, life-threatening bleeding, or death from any cause. A major adverse cardiac event was defined as myocardial infarction, SSE, or death due to a cardiovascular cause (including bleeding). All primary and secondary efficacy and safety end points were adjudicated by an independent clinical end point committee. The current analysis includes patients who represent the population for whom edoxaban was approved for clinical use by the FDA and is referred to as the FDA-approved cohort (Figure 1). This included patients randomized to higher-dose edoxaban regimen (60/30 mg) or warfarin and with a CrCl of 15 to 95 mL/min, as calculated using the Cockcroft-Gault equation 9 based on the serum concentration of creatinine measured at randomization and actual body weight. We have excluded all patients who were randomized to the lower-dose edoxaban regimen (30/15 mg, n = 7,034) because approval of this regimen was not requested by the sponsor. In addition, we excluded 1,561 patients whose dose was reduced for indications that were not included in the FDA-approved cohort (Figure 1). The ENGAGE AF–TIMI 48 trial was funded by Daiichi Sankyo, Inc, Parsippany, NJ. The authors are solely responsible for the design and conduct of this analysis, the drafting and editing of the manuscript, and its final content.

Statistical analysis Baseline characteristics were compared between groups based on renal function and treatment allocation using χ2 tests for categorical variables and Kruskal-Wallis tests for continuous variables. The primary efficacy analysis was performed with the use of a Cox proportional hazards model in the intention-to-treat population, including all events between randomization and end of the study treatment period, whether occurring on or off study drug. The primary efficacy analysis was also examined in the modified-intention-to-treat population (at least 1 dose of study drug taken) during the on-treatment period (while on drug + 3 days after the last dose or at the end of the double-blind therapy, whichever came first). All secondary efficacy analyses were performed in the intention-to-treat population. Data for safety outcomes were from the prespecified safety cohort (patients who took at least 1 dose of the assigned study drug) during the treatment period, with interval censoring of events during study-drug interruptions that lasted more than 3 days. Results are presented as hazard ratios (HRs) with 95% CIs, with P values presented without adjustment for multiple comparisons.

Results A total of 14,071 patients in the ENGAGE AF–TIMI 48 trial were randomized to either the higher-dose edoxaban

American Heart Journal Month 2016

146 Eisen et al

Figure 1

n

n

n

Study consort diagram. Patients were excluded from the current analysis (which best represent the population for whom edoxaban was approved for clinical use by the US FDA) if they were randomized to the lower-dose edoxaban regimen (30/15 mg), had a calculated CrCl N95 mL/min, or were dose adjusted for indications that were not included in the approved FDA drug label for edoxaban (Body weight≤ 60kg- 496 patients; Use of potent P-glycoprotein inhibitors- 206 patients; Body weight≤ 60kg and use of potent P-glycoprotein inhibitors- 29 patients; CrCl>50 ml/min at randomization who were dose adjusted- 399 patients; CrCl≤ 50 ml/min at randomization who were not dose adjusted- 431 patients).

regimen (60/30 mg) or warfarin. Of these 10,948 patients (77.8%) had a CrCl ≤95 mL/min at randomization. A total of 1,561 patients who were randomized to receive a reduced dose were not included in the FDA-approved cohort (Figure 1), leaving 9,387 patients included in this analysis. As compared with the 4,684 patients who constitute the nonapproved cohort (Table I), patients in the FDA-approved cohort were older, were with more than twice as many patients aged ≥75 years (49.6% vs 21.5%, P b .001), were more frequently female, and had higher CHADS2 and HAS-BLED scores. They were also more likely to have prior stroke/transient ischemic attack, but were less likely to have heart failure, diabetes mellitus, or to been from Eastern Europe. Their median CrCl was 65.2 mL/min (interquartile range [IQR] 50.3-78.1), as compared with 103.6 mL/min (IQR 62.0-122.7) in the nonapproved cohort. In the FDA-approved cohort, 86 patients (b1%) had a CrCl 15-29 mL/min at randomization and 792 patients (8.4%) had at least 1 measurement of CrCl 15-29 mL/min during the trial. In the FDA-approved cohort, 4,633 patients were randomized to the edoxaban regimen; 3,489 of these (75.3%) received 60 mg and 1,144 of these (24.7%) received 30 mg according to protocol. In addition, 4,754 patients were randomized to warfarin; their median TTR was 68.8%

(IQR 56.6-77.5). The clinical and laboratory characteristics of the patients at baseline were similar between both treatment groups (online Appendix Supplementary Table SI). The median follow-up was 2.8 years.

Efficacy end points In the intention-to-treat population, SSE occurred in 202 patients (1.63%/y) in the edoxaban group as compared with 253 patients (2.02%/y) in the warfarin group (HR 0.81, 95% CI 0.67-0.97, P = .023) (Figure 2A). Patients in the edoxaban group had fewer strokes than did patients in the warfarin group (1.55%/y vs 1.88%/y; HR 0.82, 95% CI 0.68-1.00, P = .046). This difference was driven primarily by a significant reduction in hemorrhagic stroke with edoxaban (HR 0.47, 95% CI 0.31-0.72, P b .001). Ischemic stroke rates were similar between the treatment groups (1.31%/y vs 1.39%/y, HR 0.94, 95% CI 0.76-1.16, P = .97). Key secondary end points also were decreased significantly with edoxaban as compared with warfarin (Table II). Importantly, edoxaban reduced the rate of cardiovascular death (HR 0.84, 95% CI 0.73-0.97, P = .015) and showed a trend for the reduction of the rate of death from any cause (HR 0.90, 95% CI 0.81-1.01, P = .084). While on treatment, the primary end point SSE occurred in 110 patients (1.11%/y) in the edoxaban group as

American Heart Journal Volume 172, Number 0

Eisen et al 147

Table I. Baseline characteristics of the US FDA-approved vs not approved cohorts Characteristic Age (y) Median (IQR) Female sex Region North America Latin America Western Europe Eastern Europe Asia-Pacific region and South Africa Type of AF Paroxysmal Persistent Permanent Qualifying risk factor Congestive heart failure Hypertension requiring treatment Age ≥75 y Diabetes mellitus Prior stroke or transient ischemic attack CHADS2 score 4-6 Mean (SD) CHA2DS2-VASc score 4-9 Mean (SD) HAS-BLED score ≥3 Mean (SD) D o s e re d uc t i o n a t randomization CrCl ≤50 mL/min Body weight ≤60 kg Potent P-glycoprotein inhibitor VKA experienced Medication at time of randomization Aspirin Thienopyridine Amiodarone Digitalis CrCl (mL/min) Median (IQR)

Approved cohort (n = 9387)

Nonapproved Cohort (n = 4684)

74.0 (68.0-79.0) 3775 (40.2)

65.0 (59.0-73.0) 1535 (32.8)

2090 1280 1516 2962 1539

(22.3) (13.6) (16.1) (31.6) (16.4)

1031 (22.0) 494 (10.5) 641 (13.7) 1802 (38.5) 716 (15.3)

2355 (25.1) 2175 (23.2) 4854 (51.7)

1176 (25.1) 1105 (23.6) 2402 (51.3)

5227 (55.7)

2918 (62.3)

b.001

8781 (93.5)

4398 (93.9)

.42

4659 (49.6) 3109 (33.1) 2781 (29.6)

1009 (21.5) 1971 (42.1) 1186 (25.3)

b.001 b.001 b.001

P

b.001 b.001 b.001

Safety end points The rates of major bleeding events were 3.16%/y in the edoxaban group and 3.77%/y in the warfarin group (HR 0.84, 95% CI 0.72-0.98, P = .023) (Figure 2B). Fatal bleeding, intracranial bleeding, and any life-threatening bleeding were all significantly lower in the edoxaban group (Table III). Other bleeding events were significantly lower in the edoxaban group than warfarin, except gastrointestinal bleeding which was more common with edoxaban (1.77%/y vs 1.28%/y, respectively; HR 1.39, 95% CI 1.11-1.74, P = .005). Safety end points of the nonapproved cohort are presented in online Appendix Supplementary Table SII.

.84

2349 (25.0) 2.9 (1.0)

855 (18.3) 2.7 (0.9)

b.001 b.001

7265 (77.4) 4.5 (1.4)

2647 (56.5) 3.9 (1.4)

b.001 b.001

4812 (51.3) 2.6 (0.9) 2309 (24.6)

1765 (37.7) 2.3 (1.0) 1262 (26.9)

b.001 b.001 .003

2309 (24.6) 844⁎ (9.0) 118⁎ (1.3)

399 (8.5) 541 (11.5) 347 (7.4)

b.001 b.001 b.001

5526 (58.9)

2752 (58.8)

.90

2773 233 1175 2725

1389 (29.7) 105 (2.2) 518 (11.1) 1529 (32.6)

.88 .38 .01 b.001

(29.5) (2.5) (12.5) (29.0)

65.2 (50.3-78.1) 103.6 (62.0-122.7) b.001

Data shown are n (%) unless otherwise indicated. VKA experienced, see online Appendix Supplementary Figure S1; CHADS2 score, see Gage et al15; CHA2DS2-VASc score, see Lip et al16; HAS-BLED score, see Lip et al.17 ⁎ These patients also had a CrCl ≤50 mL/min.

Net clinical outcomes Edoxaban achieved statistically significant superior net clinical outcomes as compared with warfarin (relative reductions of 9%-14%), including the prespecified composite of the primary efficacy (SSE), principal safety (major bleeding), and all-cause mortality (HR 0.91, 95% CI 0.83-0.99, P = .026), as well as 2 other net clinical outcomes that consisted of more severe events (Table III). Subgroups Analyses of the 21 prespecified subgroups revealed only one significant subgroup-treatment interaction with regard to the primary efficacy end point, SSE (online Appendix Supplementary Figure S1). Patients who were vitamin K antagonist (VKA)–naive had a 47% (HR 0.53, 95% CI 0.39-0.72) reduction in the primary efficacy end point when on edoxaban, whereas those who were VKA experienced had similar outcomes in the edoxaban and warfarin groups (HR 1.06, 95% CI 0.84-1.35, P-interaction b .001). Median TTR in VKA-experienced patients was 71.0% (IQR 60.7-79.1) and 64.7% (IQR 50.8-74.7) in VKA-naive patients. Importantly, no difference in the efficacy of edoxaban relative to warfarin was found in patients with CrCl ≤50 mL/min compared with those with CrCl 50-95 mL/min (P-interaction = .64). In 21 prespecified subgroup analyses of the principal safety end point, International Society on Thrombosis and Haemostasis major bleeding (online Appendix Supplementary Figure S2), significant interactions were observed in patients who had an edoxaban dose reduction at randomization (ie, CrCl ≤50 mL/min) and in patients enrolled at sites with a center-based TTR below the median (P-interaction b .05 for each). Use of edoxaban in these 2 more vulnerable groups was associated with even lower rates of bleeding compared with those with mild-normal renal function or higher center-based TTRs.

Discussion compared with 176 patients (1.72%/y) in the warfarin group (HR 0.64, 95% CI 0.51-0.81, P b .001). Efficacy end points of the nonapproved cohort are presented in Table SII.

In this analysis of the FDA-approved cohort, once-daily edoxaban 60/30 mg was superior to well-managed warfarin for the prevention of SSE in patients with AF

American Heart Journal Month 2016

148 Eisen et al

Figure 2

A

P

B

P

Kaplan-Meier curves for the primary efficacy (A) and principal safety (B) end points.

American Heart Journal Volume 172, Number 0

Eisen et al 149

Table II. Efficacy end points of the US FDA-approved cohort⁎

End point Primary end point Intention-to-treat population in the overall study period Modified intention-to-treat population in the treatment period † Stroke Hemorrhagic Ischemic Nondisabling and nonfatal ‡ Disabling or fatal ‡ Fatal Systemic embolic event Key secondary end points Stroke, systemic embolic event, or death from cardiovascular causes Major adverse cardiac event § Stroke, systemic embolic event, or death Other end points Death or intracranial hemorrhage Death or disabling stroke Death Any cause Cardiovascular causes Myocardial infarction

Edoxaban 60/30 mg vs warfarin in the FDA-approved cohort

Edoxaban 60/30 mg (n = 4633)

Warfarin (n = 4754)

n

Event rate (%/y)

n

Event rate (%/y)

202 110 192 31 162 109 87 58 10

1.63 1.11 1.55 0.25 1.31 0.88 0.70 0.46 0.08

253 176 235 67 175 138 103 69 19

2.02 1.72 1.88 0.53 1.39 1.10 0.81 0.54 0.15

0.81 0.64 0.82 0.47 0.94 0.80 0.85 0.85 0.53

(0.67-0.97) (0.51-0.81) (0.68-1.00) (0.31-0.72) (0.76-1.16) (0.62-1.03) (0.64-1.14) (0.60-1.21) (0.25-1.14)

.023 b.001 .046 b.001 .54 .079 .28 .37 .11

501 570 671

4.04 4.64 5.41

609 676 776

4.86 5.44 6.18

0.83 (0.74-0.93) 0.85 (0.76-0.95) 0.87 (0.78-0.96)

.002 .004 .008

585 575

4.66 4.58

688 652

5.42 5.11

0.86 (0.77-0.96) 0.89 (0.80-1.00)

.006 .043

556 371 91

4.39 2.93 0.73

623 446 101

4.83 3.45 0.80

0.90 (0.81-1.01) 0.84 (0.73-0.97) 0.91 (0.69-1.21)

.084 .015 .54

HR (95% CI)

P

⁎ Data are from the intention-to-treat cohort during the overall study period with 95% CIs and P values for superiority. † The analysis of the modified intention-to-treat population included data from 4615 patients in the higher-dose edoxaban regimen group and 4734 patients in the warfarin group. ‡ Stroke severity was assessed by means of the Rankin score.18 A score of 0, 1, or 2 indicated a nondisabling stroke, and a score of 3, 4, or 5 a nonfatal disabling stroke. A score of 6 indicated a fatal stroke. § A major adverse cardiac event was defined as myocardial infarction, stroke, systemic embolic event, or death due to cardiovascular cause (including bleeding).

and a CrCl ≤95 mL/min. In addition, treatment with edoxaban compared with warfarin was associated with a reduction in cardiovascular death as well as in major bleeding events, with significant reductions in hemorrhagic stroke, intracranial, and fatal bleeding events. Edoxaban and warfarin prevented ischemic stroke to a similar degree, whereas there was a significant excess in gastrointestinal bleeding with edoxaban. Overall, edoxaban provided superior net clinical outcomes as compared with warfarin, with a trend in total mortality favoring edoxaban. We believe that this study is of clinical importance because it reports for the first time the efficacy and safety of edoxaban in a population that has been approved by the FDA. By eliminating patients who were treated with low-dose edoxaban, had a CrCl N95 mL/min, or were dose adjusted due to low body weight or concomitant use of potent P-glycoprotein inhibitors, this study emphasizes the substantial differences between the FDA-approved cohort and the nonapproved cohort. Importantly, this study demonstrates the actual outcomes with edoxaban that might be expected in clinical practice in the United States. Although edoxaban 60/30 mg was noninferior to warfarin in the prevention of SSE in the ENGAGE AF–TIMI 48 trial, it was superior to warfarin in the FDA-approved cohort. In addition, consistent with the main trial,

edoxaban 60/30 mg in the FDA-approved cohort had a better safety profile as compared with warfarin. Interestingly, the FDA approval of edoxaban was not based solely on efficacy and safety data from the ENGAGE AF–TIMI 48 trial, but also on pharmacokinetic and pharmacodynamic modeling. As such, in the FDA label for edoxaban, in contrast to the actual dose reduction in the trial, no dose reduction was recommended based on low body weight or potent P-glycoprotein inhibition. Similarly, dabigatran 75 mg twice a day, had also earlier been approved by the FDA for patients with CrCl 15-30 mL/min based on drug exposure studies and not only on efficacy and safety outcomes. 10 It is noteworthy that these favorable findings for edoxaban in the FDA-approved cohort occurred in a higher-risk population, characterized by older age with a greater risk for ischemic and bleeding events, as compared with the nonapproved cohort enrolled in ENGAGE AF–TIMI 48 trial. As might be expected for such a high-risk population, 12.7% of the patients in the approved cohort died during the 2.8 years of average follow-up. Most (69%) of the deaths were due to cardiovascular causes, and edoxaban significantly reduced the rate of cardiovascular death by 16% relative to warfarin. These observations highlight the important benefits of edoxaban in the FDA-approved cohort. Edoxaban

American Heart Journal Month 2016

150 Eisen et al

Table III. Safety end points and net clinical outcomes of the US FDA-approved cohort⁎ Edoxaban 60/30 mg (n = 4615)

Major bleeding † Fatal bleeding Bleeding into a critical organ or area Overt bleeding with blood loss of ≥2 g/dL Any intracranial bleeding Fatal intracranial bleeding Gastrointestinal bleeding Upper gastrointestinal tract Lower gastrointestinal tract Bleeding in other location Life-threatening bleeding ‡ Fatal bleed and life threatening bleeding Clinically relevant nonmajor bleeding Minor bleeding Major or clinically relevant nonmajor bleeding Any overt bleeding Net clinical outcome § Primary Secondary Tertiary

Warfarin (n = 4734)

n

Event rate (%/y)

n

Event rate (%/y)

308 24 76 236 40 18 174 101 75 97 41 65 847 441 1080 1325

3.16 0.24 0.77 2.41 0.40 0.18 1.77 1.02 0.76 0.98 0.41 0.65 9.48 4.72 12.32 15.87

379 45 155 235 104 34 130 77 55 153 92 137 952 506 1221 1464

3.77 0.44 1.52 2.32 1.01 0.33 1.28 0.75 0.54 1.51 0.90 1.33 10.54 5.29 13.75 17.36

962 627 711

8.11 5.03 5.76

1071 729 832

8.90 5.77 6.67

Edoxaban 60/30 mg vs Warfarin in the FDA-approved cohort HR (95% CI) 0.84 0.55 0.51 1.04 0.40 0.55 1.39 1.36 1.41 0.65 0.46 0.49 0.90 0.89 0.90 0.92

P

(0.72-0.98) (0.34-0.90) (0.38-0.67) (0.87-1.25) (0.28-0.57) (0.31-0.97) (1.11-1.74) (1.01-1.82) (1.00-2.00) (0.51-0.84) (0.32-0.66) (0.36-0.66) (0.82-0.99) (0.79-1.02) (0.83-0.98) (0.85-0.99)

.023 .018 b.001 .68 b.001 .039 .005 .044 .051 .001 b.001 b.001 .031 .086 .011 .025

0.91 (0.83-0.99) 0.87 (0.78-0.96) 0.86 (0.78-0.95)

.026 .008 .003

⁎ Data are from the safety cohort during the treatment period (which began when the first dose of study drug was administered), with interval censoring of events during study-drug interruptions that lasted more than 3 days, except for net clinical outcomes, which are presented for the overall treatment period (which began at the time of randomization). † Major bleeding event is defined as a clinically overt bleeding event that meets at least one of the following: (a) fatal bleeding, (b) bleeding in a critical organ or area, and (c) clinically overt bleeding that causes a fall in hemoglobin level of 2.0 g/dL or more. Major bleeding events were also further subclassified as life-threatening or non–life-threatening. A life-threatening major bleed was defined as a bleeding event that is either intracranial or is associated with hemodynamic compromise requiring intervention. ‡ Life-threatening bleeding was defined as a bleeding event that is either intracranial or associated with hemodynamic compromise requiring intervention. § The primary net clinical outcome was a composite of stroke, systemic embolic event, major bleeding, or death from any cause. The secondary net clinical outcome was a composite of disabling stroke, life-threatening bleeding, or death from any cause. The tertiary net clinical outcome was an exploratory composite of stroke, systemic embolic event, life-threatening bleeding, or death from any cause.

was more effective than warfarin in the prevention of SSE, particularly in elderly and other high-risk subgroups of patients with AF. As observed in the ENGAGE AF–TIMI 48 trial, edoxaban demonstrated greater efficacy compared with warfarin in patients who were VKA-naive as compared with those who were VKA experienced. This could be explained by the higher TTR observed in the latter. 11 Bleeding is of major concern in the treatment with oral anticoagulants, in particular in the elderly. 12 Although nearly half of the patients in the FDA-approved cohort were 75 years or older and the HAS-BLED scores were high, edoxaban significantly reduced major bleeding and decreased the rate of intracranial bleeding by more than half, when compared with warfarin. Furthermore, edoxaban significantly decreased fatal bleeding events, as well as most categories of major and minor bleeding, with the exception of gastrointestinal hemorrhage. Interestingly, as observed in the subgroup analyses of the principal safety end point, edoxaban dose adjustment due to renal dysfunction improved safety further, with even greater reductions in major bleeding relative to warfarin among patients who had CrCl 15-50 mL/min. The better safety profile of edoxaban contributed to the lower cardiovascular mortality and to superior net clinical outcomes over warfarin in this population.

Although the subpopulation of patients with CrCl ≤95 mL/min is older and thus more representative of a typical population of patients with AF in clinical practice, 13 it is essential to assess the renal function of each patient prior to initiating edoxaban (as is true of all new oral anticoagulants) in order to determine treatment eligibility and whether dose reduction is necessary. As with all other new oral anticoagulant trials, renal function in the ENGAGE AF–TIMI 48 trial was based on a single creatinine measurement at baseline. Although creatinine could be subject to change, the intraindividual variation is low. 14 We have shown, not only that edoxaban was superior to warfarin for the prevention of SSE in AF patients with CrCl ≤95 mL/min, but also that major bleeding was even further reduced in those patients who had moderate renal dysfunction and who therefore received 30 mg. Thus, the dose adjustment of edoxaban according to CrCl, which was used in the ENGAGE AF–TIMI 48 trial, should be considered by clinicians as a means of achieving even better safety with edoxaban.

Limitations This study is a post hoc analysis of a clinical trial population and should be interpreted with an understanding of potential limitations. The cut-point of CrCl ≤95 mL/min was selected by the FDA based on post hoc

American Heart Journal Volume 172, Number 0

modeling data, and thus, less favorable results may have occurred due to regression to the mean with prospective reevaluation. In addition, patients whose dose was reduced due to low body weight or potent P-glycoprotein inhibitor therapy with CrCl N50 mL/min were not included in the FDA-approved population, and outcomes in these groups without dose reduction were not evaluated. Although the FDA approved edoxaban for patients with CrCl 15-95 mL/min, 878 patients with CrCl 15–29 mL/min were underrepresented in the ENGAGE AF–TIMI 48 trial.

Conclusion In the FDA-approved cohort for edoxaban, with a CrCl 15-95 mL/min and no dose reduction for low body weight or potent P-glycoprotein inhibition, treatment with edoxaban 60/30 mg once daily is superior to wellmanaged warfarin in the prevention of SSE. This dosage is associated with lower rates of bleeding, including fatal bleeding and intracranial hemorrhage; it reduces cardiovascular death and provides superior net clinical outcomes. These favorable results now expand the pharmacotherapeutic options of anticoagulant treatment in patients with AF.

Disclosures Drs Eisen, Nordio, Gogia, Awasty, and Henderson have no disclosures. Drs Mercuri and Rutman are employees of Daiichi Sankyo, which funded this trial. Dr Giugliano has served as a consultant and had received honoraria from Bristol-Myers Squibb, Janssen, Daiichi-Sankyo, Merck, and Sanofi, and grant support through his institution from Daiichi-Sankyo, Merck, Johnson & Johnson, Sanofi, and AstraZeneca. Dr Ruff reports grant support through his institution from Daiichi-Sankyo; has served as a consultant and received honoraria from Daiichi-Sankyo, Boehringer Ingelheim, and Bristol-Myers Squibb; and reports receiving grant support through his institution outside the submitted work from AstraZeneca, Eisai, and Intarcia. Dr Antman reports receiving grant support through his institution from Daiichi-Sankyo. Dr Braunwald reports grants (through the Brigham and Women's Hospital) and personal fees for lectures from Daiichi-Sankyo. He has received grants (through the Brigham and Women's Hospital) from AstraZeneca, Merck & Co, and Glaxo Smith Kline; personal fees for consultancies from Genyzme, Medicines Co, Theravance, and Sanofi-Aventis; and honoraria for lectures from Menarini and Medscape.

Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.ahj.2015.11.004.

Eisen et al 151

References 1. Giugliano RP, Ruff CT, Braunwald E, et al. Antman EM; ENGAGE AF– TIMI 48 Investigators. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2013;369(22):2093-104. 2. Ruff CT, Giugliano RP, Braunwald E, et al. Association between edoxaban dose, concentration, anti–factor Xa activity, and outcomes: an analysis of data from the randomised, double-blind ENGAGE AF– TIMI 48 trial. Lancet 2015;385(9984):2288-95. 3. Ogata K, Mendell-Harary J, Tachibana M, et al. Clinical safety, tolerability, pharmacokinetics, and pharmacodynamics of the novel factor Xa inhibitor edoxaban in healthy volunteers. J Clin Pharmacol 2010;50:743-53. 4. Salazar DE, Mendell J, Kastrissios H, et al. Modeling and simulation of edoxaban exposure and response relationships in patients with atrial fibrillation. Thromb Haemost 2012;107:925-36. 5. Lip GY, Agnelli G. Edoxaban: a focused review of its clinical pharmacology. Eur Heart J 2014;35(28):1844-55. 6. http://www.fda.gov/downloads/AdvisoryCommittees/ CommitteesMeetingMaterials/Drugs/ CardiovascularandRenalDrugsAdvisoryCommittee/UCM420704.pdf. 7. Ruff CT, Giugliano RP, Antman EM, et al. Evaluation of the novel factor Xa inhibitor edoxaban compared with warfarin in patients with atrial fibrillation: design and rationale for the Effective aNticoaGulation with factor xA next GEneration in Atrial Fibrillation– Thrombolysis In Myocardial Infarction study 48 (ENGAGE AF–TIMI 48). Am Heart J 2010;160:635-41. 8. Schulman S, Kearon C. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost 2005;3:692-4. 9. Cockroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31-41. 10. Beasley BN, Unger EF, Temple R. Anticoagulant options—why the FDA approved a higher but not a lower dose of dabigatran. N Engl J Med 2011;364(19):1788-90. 11. O'Donoghue ML, Ruff CT, Giugliano RP, et al. Edoxaban vs. warfarin in vitamin K antagonist experienced and naive patients with atrial fibrillation. Eur Heart J 2015;36(23):1470-7. 12. Barco S, Cheung YW, Eikelboom JW, et al. New oral anticoagulants in elderly patients. Best Pract Res Clin Haematol 2013;26(2): 215-24. 13. Iguchi Y, Kimura K, Kobayashi K, et al. Relation of atrial fibrillation to glomerular filtration rate. Am J Cardiol 2008;102(8):1056-9. 14. Reinhard M, Erlandsen EJ, Randers E. Biological variation of cystatin C and creatinine. Scand J Clin Lab Invest 2009;69(8): 831-6. 15. Gage BF, Waterman AD, Shannon W, et al. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001;285:2864-70. 16. Lip GY, Nieuwlaat R, Pisters R, et al. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor–based approach: the Euro Heart Survey on Atrial Fibrillation. Chest 2010;137(2):263-72. 17. Lip GY, Frison L, Halperin JL, et al. Comparative validation of a novel risk score for predicting bleeding risk in anticoagulated patients with atrial fibrillation: the HAS-BLED (Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile INR, Elderly, Drugs/Alcohol Concomitantly) score. J Am Coll Cardiol 2011;57: 173-80. 18. van Swieten JC, Koudstaal PJ, Visser MC, et al. Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988;19:604-7.