Before FDA could approve Daiichi Sankyo Inc.’s Factor Xa inhibitor Savaysa (edoxaban) for stroke risk reduction in atrial fibrillation patients, the agency had to overcome numerous differences of opinion – both internally and among its external advisors – on some unexpected efficacy findings in the pivotal trial.

Agency staff within and across review disciplines were divided on what to do about a significant interaction between edoxaban efficacy and renal function in a-fib patients. Specifically, patients with normal renal function fared worse than those with mild renal impairment in the Phase III ENGAGE AF-TIMI 48 trial.

Our New Drug Review Profile Series

For the third Monday of every month, "The Pink Sheet" will analyze the FDA review documents for a particular approval, examining the issues that came up during the review, how they were resolved and regulatory trends and lessons that can help all sponsors.

With our April Drug Review Profile, we found that between the limited scope of edoxaban’s approval and the culmination of issues that have affected all of the novel oral anticoagulants, there are plenty of insights to be gleaned from the fourth NOAC approval. In this issue we'll present stories on dosing, the potential need for blood monitoring with NOACs and how FDA resolves differing internal opinions. Check back next week for additional stories on edoxaban’s lost indication, the benefits of being late in class, and a 70-person, mid-cycle meeting.

Similarly, patients with moderate renal impairment, whose edoxaban doses were reduced per the trial protocol, fared worse than patients whose renal function was only mildly impaired but who received the full 60 mg dose.

Agency reviewers seemed uniformly to agree that the efficacy differences resulted from insufficient drug exposure levels in patients with normal and moderately impaired renal function. However, opinions differed as to whether edoxaban should be approved, and at what dose, for patients with normal renal function.

In a Jan. 8 memo, Office of Drug Evaluation I Acting Deputy Director Robert Temple summed up the quandary facing the agency.

“There appear to be two relatively unusual choices: 1) approval with a highest recommended dose of 60 mg, but with labeling that indicates that edoxaban should not be used in patients with normal renal function, or 2) approval with a recommended dose for those with normal renal function that is greater than 60 mg, the highest dose studied in ENGAGE AF,” Temple said.

The agency ultimately opted for the former approach, but only after identifying a threshold level of renal function at which the drug’s efficacy was adversely affected ([A#14150109005]).

Savaysa’s A-Fib Label

Indications and Usage:

Savaysa is indicated to reduce the risk of stroke and systemic embolism (SE) in patients with nonvalvular atrial fibrillation (NVAF).
Limitation of use for NVAF: Savaysa should not be used in patients with CrCL >95 mL/min because of an increased risk of ischemic stroke compared to warfarin.

Boxed Warning:

Savaysa should not be used in patients with CrCL >95 mL/min. In the ENGAGE AF-TIMI 48 study, nonvalvular atrial fibrillation patients with CrCL >95 mL/min had an increased rate of ischemic stroke with Savaysa 60 mg once daily compared to patients treated with warfarin. In these patients another anticoagulant should be used.

The disagreements over how to handle the renal subgroup efficacy findings also led to differences of opinion on the need for a Risk Evaluation and Mitigation Strategy for the a-fib indication.

FDA’s review of edoxaban for a-fib and a second indication – treatment of deep vein thrombosis and pulmonary embolism – spanned two review divisions, and the renal subgroup findings necessitated an advisory committee meeting that was not expected at the review’s outset.

Despite these complexities and the massive size of the NDA, the anticoagulant was approved on its Jan. 8 user fee goal date for both indications, reflecting the agency’s ability to reach consensus on complicated issues that can arise even for drugs that are late in class.

No Concerns, At First

Edoxaban is the fourth novel oral anticoagulant (NOAC) approved in the U.S. as an alternative to warfarin and the inconvenient blood coagulation monitoring required for the vitamin K antagonist.

Edoxaban followed in the well-worn path first traveled by Boehringer Ingelheim GMBH’s direct thrombin inhibitor Pradaxa (dabigatran), the first NOAC approved in October 2010. This was followed by the approval of two Factor Xa inhibitors: Janssen Pharmaceuticals Inc./Bayer HealthCare AG’s Xarelto (rivaroxaban) in July 2011 and Bristol-Myers Squibb Co./Pfizer Inc.’s Eliquis (apixaban) in December 2012.

Daiichi submitted the edoxaban NDA in January 2014 for multiple indications, including reducing the risk of stroke and systemic embolism in patients with nonvalvular a-fib and treatment of DVT/PE. A third requested indication – for prevention of recurrence of symptomatic venous thromboembolism – did not make its way into labeling.

[Sidebar#00150302016]

The NDA was administratively divided for review by two different divisions, with the a-fib indication going to the Division of Cardiovascular and Renal Products, and the DVT/PE treatment and VTE prevention indications going to the Division of Hematology Products (see sidebar).

The a-fib application was supported by one pivotal trial. ENGAGE AF-TIMI 48 was a double-blind, double-dummy, parallel-group study of more than 21,000 subjects. Warfarin was used as an active control, and two doses of edoxaban were tested: 60 mg, with dose adjusted to 30 mg for patients who met certain criteria (such as creatinine clearance 50 mL/min); and 30 mg, dose adjusted to 15 mg under the same criteria.

Both edoxaban doses were non-inferior to warfarin on the primary endpoint, which was the occurrence of first stroke (either ischemic or hemorrhagic) or systemic embolic event during treatment or within three days from the last dose taken.

FDA approved a highest recommended dose of 60 mg only after attempting to clearly identify a threshold level of renal function at which efficacy appeared to be adversely affected.

The 60 mg edoxaban dose came close to demonstrating superiority to warfarin, while the 30 mg dose was numerically less effective than warfarin on the primary endpoint and markedly inferior on ischemic stroke. As a result, Daiichi proposed to market only the 60 mg dose (dose-adjusted to 30 mg) for a-fib – a decision with which FDA agreed.

“Overall, the efficacy findings support approval of the edoxaban 60 mg regimen,” Cross-Discipline Team Leader Martin Rose said in a Dec. 8 review.

Nevertheless, there was one issue that stood in the way of FDA approval in the a-fib setting.

An Efficacy Hurdle Arises

Early in its review, the cardio-renal division identified an interaction between baseline renal function and efficacy, both with regard to the primary endpoint and ischemic stroke.

The most favorable results for edoxaban compared to warfarin were observed in the subgroup of patients with mildly impaired renal function in the 60 mg arm. Mild impairment was defined as estimated creatinine clearance of >50 mL/min but <80>

In contrast to the good results in the mild renal impairment group, patients with normal renal function (CrCL 80 mL/min or higher) in the 60 mg arm had a reversal of the hazard ratio versus warfarin for the primary endpoint, with results numerically favoring warfarin “by a substantial margin,” Rose said.

The normal renal function subgroup accounted for about one-third of the patients in ENGAGE and had an increased rate of ischemic stroke and total stroke with edoxaban compared to warfarin.

The subgroup of patients with moderate renal impairment (CrCL of 30 mL/min to <50 ml/min)="" also="" fared="" worse="" than="" the="" mildly="" impaired="" subgroup,="" but="" better="" than="" the="" normal="" renal="" function="" subgroup.="" about="" 80%="" of="" patients="" with="" moderate="" renal="" impairment="" were="" dose-reduced="" pursuant="" to="" the="" trial="">

Cardio-renal division clinical reviewers Melanie Blank and Tzu-Yun McDowell concluded that the poor performance in the normal renal function subgroup resulted from lower drug exposures. They also pointed to a mechanistic basis for their findings – edoxaban is 50% renally excreted.

The concerns about under-dosing in a-fib patients with normal renal function did not extend into the DVT/PE setting (see related story, Also see "Savaysa Under-Dosing Not A Concern In DVT/PE Treatment" - Pink Sheet, 20 April, 2015.).

Blank and McDowell identified several options for addressing the renal function-related efficacy findings for the a-fib indication.

The agency could issue a “complete response” letter and require a new efficacy trial, preferably with a dose greater than 60 mg, in the normal renal function subgroup. Or, it could approve edoxaban 60 mg but only for patients with renal impairment.

A third option was to approve a dose greater than 60 mg for patients with normal renal function. Pharmacometric modeling would be used to determine a dose that would result in drug exposure levels similar to those achieved by patients with mild renal impairment who received the 60 mg dose in ENGAGE.

FDA decided to convene its Cardiovascular and Renal Drugs Advisory Committee on Oct. 30 to discuss the options. Agency reviewers’ presentations to the committee reflected the divergence of opinion on what to do with the application.

Blank and McDowell recommended approval of the 60 mg dose with a limitation of use to patients with abnormal renal function (CrCL <80 ml/min)="">[A#14141028001]).

FDA’s statistical reviewer recommended approval of the 60 mg dose for most patients, including those with normal renal function. In contrast, the clinical pharmacology reviewer supported approval in normal renal function patients of a higher-than-tested dose based upon pharmacometric modeling.

The committee voted 9-1 in favor of approval, but it too was divided on the options regarding dose and patient population.

Five panelists favored approval of the 60 mg dose for patients with normal or mildly impaired renal function. Two panelists said the drug should be approved at a dose higher than 60 mg for patients with normal renal function, and two said the 60 mg dose should be approved only for patients with mild and moderate renal impairment.

At the meeting’s close, FDA Division of Cardiovascular and Renal Drug Products Director Norman Stockbridge said the panel “achieved about the same level of consensus that we achieved internally” ([A#14141030008]).

What To Do?

The thinking of FDA reviewers continued to evolve after the advisory committee meeting.

In his cross-discipline review, Rose noted “the majority of review team members, as well as the office signatory and division director, believe that it is both feasible and appropriate to use PK modeling and exposure matching to determine a dose” for normal renal function patients.

Yet, he also acknowledged the clinical reviewers’ initial concerns that “there are too many unknowns and too much sensitivity in the medical community regarding bleeding risk to approve a dose higher than any dose use in Phase III.” Bleeding is the key safety drawback with anticoagulants.

However, Rose said the ENGAGE data suggest that increasing the amount of unabsorbed edoxaban would not have large effects on gastrointestinal bleeding, which was one of the primary safety concerns with the drug.

Although Rose’s approval recommendations are partially redacted from his publicly released memo, it’s clear he supported approval of a dose greater than 60 mg for patients with normal renal function, with “prominent language” in labeling regarding the renal function subgroup findings.

In a Dec. 23 memo, Stockbridge also favored approval of a higher-than-tested dose, saying such an approach is not entirely without precedent.

“Where the relationship between exposure and effect matters and the factors affecting exposure are understood, dose adjustments are common. Although the majority of such adjustments result in a decrease in dose, upwards adjustment to compensate for the effects of metabolic inducers is not rare,” Stockbridge said.

There is “discomfort with approval of a drug intended for wide use in a relatively fragile population at a never studied dose when there are available alternative NOACs,” FDA’s Temple said.

“I would label with dose recommendations of 75 or 90 mg for patients with creatinine clearance over 80 mL/min, 60 or 75 mg for those with creatinine clearance 50-80 mL/min, and 30 or 45 mg for patients with creatinine clearance <50>

In his memo, Temple noted that a “wide range of views have been expressed by reviewers, the [cross-discipline team leader], advisory committee members, and at internal meetings by FDA management. Virtually all have believed that edoxaban should be approved, but with diverse views on just how to do that.”

Given what is known about edoxaban plasma concentrations and their effect on stroke rate with acceptable bleeding, “a very attractive option is approval of a dose larger than the 60 mg dose studied in ENGAGE AF, an option supported by Clinical Pharmacology and Drs. Rose and Stockbridge,” Temple said.

Clinical pharmacology reviewers suggested a dose of approximately 75 mg-90 mg based on the extensive concentration-response data available, Temple noted. “These doses have, of course, not been actually studied to any extent and the advisory committee, except for two members, did not support such a step.”

Temple acknowledged there is “discomfort with approval of a drug intended for wide use in a relatively fragile population at a never studied dose when there are available alternative NOACs. I do, however, believe the promising results in the mild renal failure group make availability of edoxaban desirable, even if labeling limits use to patients with a defined level of renal function.”

After the advisory committee meeting, the cardio-renal clinical reviewers performed further efficacy analyses by creatinine clearance rates to better define a population in which edoxaban 60 mg performed well.

It “seems clear from numerous decile and quintile assessments, that although effect on stroke clearly diminishes overall with increasing renal function, the place at which it becomes worse than warfarin is not perfectly clear. It does, however, appear adverse in the patients with CrCL of above 98 mL/min.,” Temple said.

“It is noted that although the best results seem to be in patients with clearly, but mildly, reduced renal function (CrCL <80 ml/min),="" results="" in="" the="">80-95 mL/min group appear similar to warfarin. We are therefore approving use of the 60 mg dose in patients with CrCL up to 95 mL/min, and recommending use of an alternative drug in people with better renal function,” Temple said.

Temple also called for blood level monitoring of NOACs to optimize their use (see related story, Also see "Novel Anticoagulants Need Blood Level Monitoring Measures, FDA’s Temple Says" - Pink Sheet, 20 April, 2015.).

Is A REMS Necessary?

In their initial recommendation that the 60 mg dose be approved only for patients with abnormal renal function, Blank and McDowell strongly argued that a REMS with elements to assure safe use – such as a restricted distribution system that incorporated mandatory creatinine clearance testing – would be needed to ensure the drug was not used in patients with normal renal function.

They recognized, however, that this type of risk management strategy could be problematic.

“Such a REMS would have a negative impact on use of the drug by making it difficult to prescribe,” the reviewers said. “Also, it would be difficult to fashion an ETASU REMS for the atrial fibrillation indication if edoxaban is approved for treatment of deep vein thrombosis at a 60 mg dose without a REMS.”

However, FDA’s Division of Risk Management saw no need for a REMS even though labeling recommended against use in a-fib patients with higher renal function.

If edoxaban is used as indicated in patients with creatinine clearance below a certain threshold, additional risk mitigation measures beyond labeling are not necessary, risk management analyst Cathy Miller said in a Jan. 5 review. If edoxaban is used in patients with creatinine clearance levels above the identified threshold, patients will be subject to lack of efficacy rather than increased risk of adverse events, Miller said.

Savaysa A-Fib Data Led To Conflicting FDA Views On Dose, Population

More from United States

More from North America