Safety Concerns Turn FDA Panel Thumbs Down for Novel CKD-Anemia Drug

An FDA advisory committee almost unanimously agreed that the risks were just too high for the chronic kidney disease (CKD)-related anemia drug roxadustat in any patient population.

In a 13-1 vote Thursday, members of the Cardiovascular and Renal Drugs Advisory Committee felt that although there’s an unmet need for new anemia therapeutics, this particular drug carried too strong of a safety signal for thrombotic risk in a non-dialysis dependent patient population.

Panel voting members felt similarly when asked about whether roxadustat should be approved specifically for patients on dialysis, voting 12-2 against recommendation for approval.

“This was a challenging vote for me,” said Thomas J. Wang, MD, of University of Texas Southwestern Medical Center in Dallas, who voted ‘no’ on both populations. “I do appreciate the potential benefits of this medication, as well as the unmet clinical need — especially for the ESA-nonresponsive patients.”

“I also appreciate the potential benefits of this novel mechanism of action,” Wang added. “My hope is that the applicant is going to be able to attain further data regarding the benefits and the risks to provide reassurance about some of the concerns that’ve been raised so that this medication can move forward.”

The agent’s sponsor, FibroGen in partnership with AstraZeneca and Astellas, developed what could be the first-in-class oral anemia treatment, as the only other available treatment options for anemia include iron, erythropoiesis stimulating agents (ESAs), and red blood cell transfusions.

Roxadustat acts as a reversible inhibitor of hypoxia inducible factor-prolyl hydroxylases (HIF-PH), orally administered and titrated up to achieve a target hemoglobin level. The sponsor was seeking an indication for the treatment of anemia due to CKD in adults either on or not on dialysis.

However, in the six-study clinical trial program — three in a non-dialysis dependent patient population plus another three in a dialysis-dependent population — a clear safety signal appeared, despite the treatment demonstrating efficacy. Specifically, there was an increase in the incidence of thrombotic events.

“I do not think that the rise in hemoglobin is the cause of thrombosis,” said Leslie S. Cho, MD, of the Cleveland Clinic in Ohio, who voted ‘no’ on for both populations. Instead, she suggested that roxadustat’s novel mechanism of action — HIF-inhibition — may be at the root of the thrombosis risk.

In order to offset the risk of thrombosis, which appeared to increase when hemoglobin levels rose at a higher rate, the sponsor suggested a different dosing protocol than was followed in the phase III program. During their presentation to the FDA panel, FibroGen instead suggested aiming for a target hemoglobin level between 10.0 to 11.0 g/dL, instead of the 10.5 to 12.0 g/dL that was used in the trials. They also proposed lower starting doses of roxadustat, nearly half of what was followed in the phase III dosing, with a cap of no more than 3 consecutive dose increases in non-responders.

Despite these new proposals, panel members weren’t convinced that this would ultimately offset the apparent risks, while maintaining efficacy.

“I think these safety concerns appear quite real,” said C. Noel Bairey Merz, MD, of Cedars-Sinai Medical Center in Los Angeles, who voted ‘no’ twice. “I don’t think the FDA should approve on the basis of mitigation modeling.”

Edward K. Kasper, MD, of Johns Hopkins School of Medicine in Baltimore, echoed this feeling when justifying his two ‘no’ votes: “I would like to see the mitigation strategy tested prior to a re-vote.”

Despite an overwhelming number of negative votes, many panel members expressed their disappointment in doing so, underscoring the great burden that these patients carry.

During an emotional open public hearing forum, many CKD patients voiced why new therapeutic options are desperately needed, particularly because many were fearful over developing antibodies from red blood cell transfusions, which could potentially make them a non-match for a donor kidney.

Being the only panel member to vote “yes” to both questions, Susan T. Crowley, MD, MBA, of Yale University in New Haven, Connecticut, said, “I think that with the significant patient burden of anemia and the [associated] risks with any ESA that are similar in roxadustat, that a fair way potentially to do this is to allow patients and providers to do shared decision-making under a [Risk Evaluation and Mitigation Strategy] program to try to mitigate some of the risks.”

During a post-vote panel discussion, many members seemed piqued by the idea of roxadustat serving just a dialysis-dependent patient population who are erythropoietin (EPO)-resistant. In fact, several members recommended the sponsor carry out an additional study using their proposed lower-dosing strategy in this particular population with few other options.

“The real interesting opportunity here is a randomized trial of EPO and roxadustat in people who are EPO-resistant,” suggested Milton Packer, MD, of Baylor University Medical Center in Dallas, who also voted ‘no’ for both. “The real wonderful opportunity here is that the sponsor could test its lower dosing strategy … and they could even show that the improvement in hemoglobin results in an improvement in quality of life or other benefits that one would expect with an improvement in hemoglobin.”

This particular measure was a sticking point for many panel members, who wished the sponsor had quantified quality of life data for these patients.

“It would be so straight-forward to then assess benefit-to-risk in a defined patient population,” Packer added. “The EPO-resistant group seems to be a real potential for future research.”

Roxadustat is already approved in China and Japan for dialysis-dependent and non-dialysis dependent patients. Just last month, a European Medicines Agency advisory committee voted in favor of the agent for marketing authorization.

Although the FDA is not required to follow its advisory committees’ recommendations, it typically does.