Stealth’s advisory committee meeting showcases power of patient voice

Barth syndrome advocates showed how tenacious patients can seize their destiny

Following a dramatic advisory committee meeting on elamipretide from Stealth BioTherapeutics, FDA finds itself in a familiar, uncomfortable place. Advocates of patients with a devastating ultrarare condition with no approved treatments are demanding that the agency overlook flawed data to approve a drug, while the agency’s reviewers are adamant the data do not support approval.

All parties agree on one thing — it could be impossible to generate conclusive data.

The scenario makes elamipretide’s review a case study in a difficult question that has been on display in many rare disease drug reviews, and surely will be again. In the context of urgent unmet medical needs, when should powerful anecdotal evidence and imperfect efficacy data from very small trials tip the balance in favor approval? And how can FDA ensure that people with very rare diseases get access to therapies, even if it is impossible to obtain the kind of definitive evidence that regulators typically require?

At an Oct. 10 meeting, the Cardiovascular and Renal Drugs Advisory Committee gave comfort to Barth syndrome patients and caregivers, voting 10-6 that Stealth BioTherapeutics Inc. had demonstrated the efficacy of elamipretide.

It is not clear, however, whether explanations committee members provided for their votes will lead to an approval decision.  

Whatever action the agency takes, the story of the tiny biotech’s efforts to create a treatment for a disease with only 130 diagnosed males in the U.S., and about 300 worldwide, offers lessons for biopharma companies, regulators and patient advocates.

The 18-year attempt to transform elamipretide from a serendipitous scientific discovery into an FDA-approved treatment for Barth syndrome is a testament to the power of organized patients and caregivers to shape their own destiny.

At the same time, it shows the tenacity required for drug developers to overcome scientific, regulatory and financial obstacles to creating therapies for ultrarare diseases.

Elamipretide has been shuffled through four FDA divisions, received a refuse-to-file letter, and been the subject of varying and mutually exclusive recommendations from agency officials. The molecule’s tangled path through FDA illuminates the mismatch between a regulatory system that was created at a time when no one would have contemplated creating a drug for less than a thousand people and the realities of creating therapies for ultrarare diseases.

Like other recent rare disease regulatory controversies, the elamipretide odyssey is a reminder of the confusion over whether and how FDA should apply regulatory flexibility, and of the gaps between top FDA officials and reviewers over the standards that should be used to assess ultrarare disease therapies.

It is unlikely that Stealth would have had the opportunity to present elamipretide to an advisory committee if senior FDA staff, prodded by patient advocacy and media attention, hadn’t taken an interest in the drug. And it is virtually certain that a majority of the committee’s members would not have voted to endorse the drug if Patrizia Cavazzoni, director of FDA’s Center for Drug Evaluation and Research, and Peter Stein, director of CDER’s Office of New Drugs, had not attended and spoken at the advisory committee meeting.

The elamipretide experience holds lessons that could be valuable beyond FDA’s doors, including for members of Congress who are dithering over the reauthorization of rare pediatric disease priority review vouchers.

The chance to obtain a voucher worth more than $100 million made it possible for Stealth to try to create a treatment for Barth syndrome, even though the company and its investors knew they would never recoup their costs through sales of the drug. In the absence of a voucher, the company wouldn’t have taken even the first step toward helping these patients.

A dramatic advisory committee meeting

The Cardiovascular and Renal Drugs Advisory Committee meeting was a dramatic example of how external advisors can influence FDA’s deliberations, and of the impact of advocates on committee deliberations.

In the company’s presentation, it parted the curtain on a drama that started in 2006, when Hazel Szeto, a pharmacologist at Weill Cornell Medical College, discovered that elamipretide, a peptide she had been studying for an entirely different reason, penetrates cells to bind cardiolipin in the inner mitochondrial membrane, increasing adenosine triphosphate (ATP) production and decreasing production of reactive oxygen species.

In 2014, the Barth Syndrome Foundation, which had been searching for potential therapies, asked Stealth, a company that had been created to exploit Szeto’s discovery, to consider developing elamipretide for their disease, which is caused by mutations in cardiolipin. By stabilizing the lipid, the hope was that elamipretide would restore the structure and organization of the membrane needed to support ATP production.

Researchers at Johns Hopkins University, one of only two multidisciplinary centers in the world treating Barth syndrome, started the Phase II/III TAZPOWER trial of elamipretide in May 2017. The 28-week double-blind, placebo-controlled crossover study enrolled 12 boys and young men at least 12 years old, and was followed by a 168-week open-label treatment extension. The primary endpoints were improvement on the six-minute walk test (6MWT) and on the BTHS Symptom Assessment (BTHS-SA) scale.

In November 2018, Stealth learned that neither primary endpoint had been met in the blinded portion of the trial. However, researchers saw what they believed were important improvements in function in the eight patients who completed the open-label extension trial. At 36 weeks, these included a 96-meter improvement in the 6MWT (p=0.024), a significant improvement in BTHS-SA (-2.1 points, p=0.031), and improvements in secondary endpoints including knee extensor strength, patient global impression of symptoms, and some cardiac parameters.

Although Stealth believed the results were compelling, FDA told the company it would need additional data.

To bolster its case for approval, Stealth proposed a natural history study, which it conducted despite FDA advising that it would not be suitable for demonstrating a treatment effect because the 6MWT is an “effort-dependent” endpoint, meaning patients might not be motivated to give their best effort if they aren’t on therapy, and because data from an external control group could not counter the negative data from the randomized, placebo-controlled trial.

FDA and Stealth could not agree on the design of a trial that was feasible and capable of producing data that could lead to an approval.

Stealth went ahead with the natural history study, working closely with the Barth Syndrome Foundation. The foundation, which is in touch with about 90% of Barth syndrome patients worldwide, holds biennial meetings for the patient community. At these events, patients receive assessments, including tests that are identical to those conducted in Stealth’s clinical trial. Those assessments were used to develop a natural history database of 79 patients, over 50% of the entire U.S. Barth syndrome population. From this database, 19 patients were selected who matched the elamipretide clinical studies.

According to the company, patients who received elamipretide for up to one year had a mean improvement of greater than 80 meters on their 6MWT, versus less than one meter for subjects in the natural history group.

Kate McCurdy, chair of the Barth Syndrome Foundation, and the mother of a boy who died from the disease, told the advisory committee that elamipretide had led to a “transformational reduction in daily symptoms,” and had given boys and young men the ability for the first time in their lives to attend a full day of school or to work full-time.

Skeptical reviewers and enthusiastic patients

Following Stealth’s presentation, FDA reviewers spent 90 minutes shredding the company’s evidence.

The dramatic jump in 6MWT results occurred at the beginning of the open-label study, suggesting that the improvement in the effort-dependent variable was the result of unblinding, FDA said.

The agency called into question the strength of Stealth’s preclinical data, and deemed the natural history study uninterpretable, pointing to concerns about selection bias. In addition, because patients in the natural history study knew they weren’t receiving active treatment, they may not have tried as hard as patients in the open-label study, FDA staff said.

Agency staff said repeatedly that while FDA can exercise regulatory flexibility when considering therapies for very small populations, sponsors must provide at least one adequate and well-controlled study supported by confirmatory evidence.

By the time the meeting broke for lunch, it looked like there was no way the advisory committee would conclude that Stealth had conducted an adequate and well-controlled study.

After lunch, focus turned from the front of the room, where the committee and FDA officials sat, to the back of the room where most of the seats were filled with 40 people wearing blue t-shirts identifying them as Barth Syndrome Foundation advocates.

A dozen individuals who had been selected to participate in the open public hearing, all advocates, spoke from the podium, and three addressed the committee virtually.

Young men said elamipretide had rescued them from debilitating effects of Barth syndrome, and that losing access to the drug would destroy their lives.

Parents said the drug had restored the health of babies whose prognoses had been grim.

Physicians reported that the effects of elamipretide were real, without precedent, and could not be attributed to a placebo effect.

Some of the most compelling public testimony came from parents and physicians who reported on five babies who received elamipretide through expanded access requests. Three infants with Barth syndrome who had been scheduled for heart transplants improved so much that they no longer need transplants, a trajectory that physicians said they had never seen before.

In another case, elamipretide seems to have kept a baby near death alive long enough to receive a transplant.

The fifth baby, who had significant medical conditions in addition to Barth syndrome, died.

Although no one under 12 years-old was included in Stealth’s trials, the anecdotes about infants are important because about half of the deaths from Barth syndrome occur in the first year of life, and 85% of deaths attributed to the disease are in children under five.

By the time the one-hour open public hearing was finished, the mood in the room had changed. It now seemed unlikely that a majority of advisory committee members would want to stand in the way of patients gaining access to elamipretide.

It wasn’t clear, however, how they could justify voting to affirm that Stealth had demonstrated elamipretide’s efficacy given FDA’s assertion that an adequate, well-controlled study was a prerequisite for approval.

Several committee members said they were struggling to reconcile their belief, influenced by the public testimony, that Barth syndrome patients should have access to elamipretide, with statements from Hylton Joffe, director of the Office of Cardiology, Hematology, Endocrinology, and Nephrology, and presentations from FDA reviewers, strongly suggesting that Stealth had not cleared the bar for approval.

Cavazzoni and Stein step in

Interventions from Cavazzoni and Stein essentially gave the committee permission to integrate comments from the public hearing into their thinking, to be persuaded by the natural history study, and to adjust their expectations for rigorous evidence based on the small size of the population.

Cavazzoni noted that FDA can exercise regulatory flexibility when reviewing treatments for rare conditions. “In a disease such as this one, an ultrarare disease, we accept a greater degree of residual uncertainty compared to a high-prevalence disorder,” she told the committee.

Although she only spoke once, it is likely that committee members recognized that the presence of a center director at an advisory committee meeting is unusual.

In determining whether the evidence meets the legal requirements for approval, Stein said the failed placebo-controlled study should not necessarily trump other evidence. A failed randomized placebo-controlled trial does not mean that a drug is ineffective, he said.

“There are very good explanations for why a trial was not designed properly to be positive and then a subsequent trial designed more properly could be positive,” Stein said.

Stein also said the committee and FDA could consider either the natural history study or the open-label extension studies adequate and well-controlled. “In the regulations there are a number of different designs that are considered potentially adequate and well-controlled trials,” including an externally controlled or a baseline-controlled trial.

Explaining the votes

Like most of the other committee members, Carole Tucker, associate dean for research at the University of Texas Medical Branch, said she had struggled with her vote. In the end, she said anecdotal evidence from the public hearing and positive trends from Stealth’s open-label study “just snuck over the line to a ‘yes’ for me.”

Explaining his yes vote, Eric Peterson vice provost at the University of Texas Southwestern Medical Center, cited evidence from the open-label extension trial and from the historic controls, noting that both datasets were flawed, but they had “a pretty big effect.” He also cited “responses from the community and from the physicians who are treating these patients.”

Jonathan Soslow, professor of pediatrics at Vanderbilt University Medicare Center, said the evidence may not have met the bar FDA usually requires, but “there needs to be a different structure for rare diseases.” In the face of a “preponderance of subjective evidence” indicating that elamipretide helped patients, the agency needs to exercise regulatory flexibility and ensure that Barth syndrome patients “get the opportunity to take this drug.”

Gerald Berry, professor of pediatrics at the Harvard Medical School, noted that he had treated Barth syndrome patients. “To deprive somebody of being able to get the medicine that might help is just untenable for me.”

Like the other committee members who voted ‘yes,’ Priya Kishnani, chief of medical genetics at Duke University, explained her decision in terms that were far from a ringing endorsement.

"These data were not completely compelling,” Kishnani said. She added that the natural history study “really did not meet the criteria” for an adequate, well-controlled study. "There was not a good natural history study. It was almost like a convenience sample."

Kishnani said she was persuaded to vote that efficacy had been demonstrated because the “patient testimonies were remarkable.”

Although Susan Ellenberg, professor emerita of biostatistics, medical ethics and health policy at the University of Pennsylvania, was also impressed by the testimonies, she doesn’t think they can support a conclusion that efficacy has been demonstrated. “The most compelling thing that we heard was from the people who spoke in the open public hearing, and it's hard to know what to do with that.”

Ellenberg noted that medical history is littered with cautionary stories of treatments being adopted based on patients being completely convinced they were effective, and then being discarded after carefully conducted trials demonstrated that they were not beneficial.  

Caleb Alexander, professor of epidemiology and medicine at Johns Hopkins Bloomberg School of Public Health, cast one of the six ‘no’ votes. “I don't think regulatory flexibility gets you beyond the paucity of affirmative evidence in this setting,” he said.

A philosophical question

Describing why he reluctantly voted that Stealth has not demonstrated the efficacy of elamipretide, Tobias Gerhard highlighted the challenges the committee members faced.

Gerhard, who directs the Center for Pharmacoepidemiology and Treatment Science at Rutgers University, said the “available evidence supporting the effectiveness is unfortunately not conclusive, even with somewhat relaxed standards. The concerns raised by FDA regarding selection bias, residual confounding, and limitations of effort-dependent endpoints are all valid and substantial.”

At the same time, he noted that the “data presented does by no means allow a conclusion of an absence of benefits. Much of the data in front of us are very compelling.”

Gerhard added that “in the context of a disease as rare as Barth Syndrome, I'm still not sure what that should mean for the approval decision given the difficulties in generating meaningful additional data in the foreseeable future and the large proportion of patients that are on the drug.”

The committee was faced, he said, with a philosophical question: “What if you had a hypothetical drug for an otherwise untreatable rare disease with a 50% probability of effectiveness that is reasonably safe, for which you will not likely get better data, and that has no current alternatives.” He concluded that “we're pretty close to this scenario.”

In the end, he said that although he voted that Stealth had not provided evidence of efficacy, “if you'd asked me about drug approval, I likely would have voted yes.”

That is precisely the question FDA will have to answer as it ponders the data and its advisory committee’s deliberations.

The elamipretide PDUFA goal is Jan. 29.