Why FDA should not grant full approval to Sarepta’s DMD gene therapy

DMD patients would be exposed to unacceptable risks, and the accelerated approval pathway would be degraded

FDA’s upcoming decision on Sarepta’s gene therapy will be a test of the integrity of the accelerated approval pathway, and more generally of the agency’s commitment to science-based regulation. The well-being of vulnerable patients, as well as public confidence in the agency, are at risk.

The Sarepta decision comes at a time when the actions of FDA’s leaders have raised concerns that the agency is abandoning the rigor that has been a cornerstone of medical progress, and that it is using “regulatory flexibility” as a cover for making idiosyncratic decisions that reflect empathy for desperate patients rather than evidence. The concerns are not new.   

FDA has already granted accelerated approvals to three exon-skipping drugs and a gene therapy from Sarepta Therapeutics Inc. (NASDAQ:SRPT) largely based on compassion for boys with Duchenne muscular dystrophy rather than science and data. Instead of basing decisions on rebuttable assumptions, the agency has repeatedly adjusted its requirements to fit the assumption that Sarepta’s products work.

Acceding to Sarepta’s request to convert the age-limited accelerated approval of Elevidys delandistrogene moxeparvovec to a full approval for all ages would create risks for DMD patients who would be precluded from receiving other gene therapies that are under development.

There could also be wider repercussions.

Patients with other fatal and serious conditions, and companies developing products that are intended to help them, will demand that FDA apply the same “regulatory flexibility” in the face of equivocal data.

Critics will jump on a decision to grant full approval to Elevidys, arguing that it shows why new barriers to accelerated approval should be erected. They will call for payers to have more discretion to deny coverage or reduce reimbursement to therapies marketed under accelerated approvals. 

Based on information in the public domain, a decision to grant full approval to Elevidys would violate a fundamental principle underlying the accelerated approval pathway, that sponsors must demonstrate that their drug provides clinical benefit. Under accelerated approval, the requirement to produce evidence of clinical efficacy is deferred, but it is not diminished.

A drug company could not expect to receive approval for a medicine intended to treat DMD based on a Phase III trial that failed to meet its primary endpoint and that achieved statistically significant but barely discernable results on secondary endpoints.

Sarepta says that the the North Star Ambulatory Assessment, the primary endpoint in the EMBARK trial that was intended to confirm clinical efficacy of Elevidys, is not sufficiently sensitive to detect benefit over a one-year period in the population that participated in the trial. The company, however, claims to be the world leader in the development of drugs to treat DMD, and it selected both the endpoint and the patient population.

Whether the results of the secondary endpoints — differences of 0.64 seconds on time to rise and 0.42 seconds for the 10-meter walk test at the one-year mark — predict future clinical benefit is debatable. On their own, however, they are too small to constitute the kind of meaningful clinical benefit that should support full approval.

It remains possible that the small treatment effect will be cumulative over time, eventually adding up to meaningful benefit. Sarepta’s hypothesis that it has detected the beginning of lasting halt in disease progression could turn out to be correct. It’s also possible, however, that the effect of the non-integrating gene therapy will diminish over time.

Everyone hopes the effects seen in the EMBARK trial are real evidence of a halt or slowing in disease progression, but even patient advocates who have invested financially and emotionally in Elevidys aren’t confident making that claim.

Although patients in the placebo group were allowed to cross over to treatment after one year, delaying the approval decision until two- or three-year data are available from the original treatment group may be the maximum “regulatory flexibility” warranted. If Sarepta’s claims are true and DMD progression is arrested, meaning, for example, that ambulatory patients never need to use a wheel chair as the company suggested to BioCentury, efficacy should become evident even without a control group. 

Short of such a profound benefit, FDA must require additional data to demonstrate clinical benefit, for example from ongoing trials evaluating the safety and efficacy of Elevidys in older and non-ambulatory patients. 

In some recent decisions, FDA seems to have been swayed by arguments that it should approve drugs for rare conditions because there are no alternative treatments and there are few or no risks associated with a therapy. That isn’t the case with Elevidys.

Patients who are exposed to Elevidys will be unable to receive any of the other DMD gene therapies that are under development. This is an immense risk in the context of a progressive, fatal disease.

The decision by Peter Marks, director of FDA’s Center for Biologics Evaluation and Research (CBER), to overrule agency reviewers and grant accelerated approval to Elevidys was a mistake. He conducted analyses that FDA would never accept from a drug sponsor, and carved out a narrow age group that even Sarepta and patient advocates say has no basis in medicine or biology. 

Granting full approval based on the data available now would compound the error, intensifying the backlash against accelerated approval that was sparked by the unjustified accelerated approval of Aduhelm aducanumab for Alzheimer’s disease.

The case for granting full approval to Elevidys is further degraded by Sarepta’s failure to demonstrate that the microdystrophin expression induced by its exon-skipping drugs leads to clinical benefit.

Sarepta’s Exondys 51 eteplirsen received accelerated approval in September 2016 based on the theory that a tiny amount of microdystrophin would produce clinically relevant benefits. More than seven years later, the company has not completed the trial needed to confirm or refute that theory.

Marks justified his decision to grant accelerated approval to Elevidys by disregarding the work of scores of CBER staff and conducting his own post hoc analyses of the still-unvalidated surrogate endpoint of microdystrophin in a tiny subpopulation. He said he was limiting the approval to 4-5-year-old patients in part because doing so would not disrupt the ongoing Phase III trial, Study 301, that was intended to demonstrate clinical benefit.

Now that the results of Study 301 are in hand, Sarepta plans to ask FDA to convert the accelerated approval of Elevidys to a full approval, and to lift all age restrictions.

When Marks informed Sarepta in May that he planned to grant accelerated approval for a restricted age range, he advised the company to modify its plan for the ongoing trial to prespecify an analysis of 4-5-year-old patients.

The minutes of a May 22 teleconference between FDA and Sarepta indicate that FDA had previously suggested, and the company had resisted powering the trial for this analysis. The meeting minutes also make it clear that failure to demonstrate statistically significant benefit on the primary endpoint, or on a prespecified subset of the primary endpoint, could cause FDA to decide against granting full approval.

“In the context of the lack of clinical evidence of benefit in the 6 and 7 year old children, as the team has discussed with the applicant previously, FDA again urged the applicant to consider modifying the ongoing Study 301 trial to be powered for demonstrating efficacy in the 4 to 5 year old subset, or to be prepared to have to conduct an additional study if study 301 fails its primary endpoint yet indicates likely efficacy in 4 and 5 year old subgroup,” the meeting minutes stated.

It isn’t clear, however, that Sarepta could persuade patients to enroll in another placebo-controlled study.

FDA told Sarepta that if the “study fails in the overall population but wins in the younger age subgroup of 4 to 5 years old, and if the applicant does not specify an inferential subgroup analysis, the applicant won’t be able to proceed with testing the subgroup effect following a failed test of the overall population.”

In its announcement of the results of Study 301, Sarepta did not mention prespecified analysis of any subset of the primary endpoint.

The teleconference meeting minutes and internal CBER meeting memos FDA has released to date do not indicate that Marks or anyone else at FDA contemplated the possibility of granting an approval based on secondary endpoints if the study failed to meet the primary endpoint or a prespecified subset of the primary endpoint.

While FDA has stated that it is theoretically possible for success in demonstrating benefit on secondary endpoints to overcome the failure of the primary endpoint, there are few if any examples of this in practice. The agency’s dogma has been that, with rare exceptions, secondary endpoints cannot be validly interpreted if the primary endpoint does not demonstrate statistical significance. Given the small effect sizes, it is difficult to sustain Sarepta’s argument that Study 301 constitutes such an exception.

Declining to grant full approval to Elevidys until additional data become available would not deprive patients of access. Until it has demonstrated clinical benefit, Sarepta would be able to continue to market it under accelerated approval to 4-5 year-olds, and could offer it on an expanded access basis to older patients. 

DMD patients and their families have made extraordinary sacrifices in the pursuit of treatments and cures. They deserve a regulatory process that forces drug companies to produce accurate, actionable evidence.

Signed commentaries do not necessarily reflect the views of BioCentury.