FDA’s different Peter principle problem: prioritizing hope over data

Approval of Elevidys is a sign of systemic problems

Everyone who cares about medical product development should be disturbed by FDA’s — actually Peter Marks’ — decision to grant full approval to Elevidys for Duchenne muscular dystrophy. The approval highlights systemic problems at the Center for Biologics Evaluation and Research and points to similar dysfunction at the Center for Drug Evaluation and Research.

To be clear, I believe that the approval is misguided, and that it contravenes the fundamental purpose of science-based decision-making, which Edwin Land defined as “a technique to keep yourself from kidding yourself.”

It seems that FDA has adopted a new Peter principle: when reviewing a proposed therapy for a terrible disease, a center director can cite any glimmer of positive data to justify approval. It isn’t an issue of promoting the incompetent, but rather of disregarding the competence of long-serving staff.

Even if CBER Director Marks turns out to have been right about the efficacy of Elevidys delandistrogene moxeparvovec from Sarepta Therapeutics Inc. (NASDAQ:SRPT), unilaterally overruling his staff is indicative of management failures. Because FDA granted full approval for ambulatory DMD patients age 6 and older, definitive evidence about efficacy in that population may never be produced. And given the company’s penchant for delay, confirmatory evidence in the accelerated approval population, non-ambulatory DMD patients, may be a very long time coming.

To be disappointed by the decision, it isn’t necessary to agree with Nicole Verdun, director of CBER’s Super Office of Therapeutic Products, Lola Fashoyin-Aje, acting director of CBER’s Office of Clinical Evaluation, and all of the staff who reviewed Sarepta’s data and found it wanting.

Commenting on data that Marks asserted are “compelling,” the CBER review team concluded that it is impossible to “reliably distinguish if these results are due to actual effects of Elevidys, or to chance alone.” They stated that “because the primary efficacy analysis was not statistically significant, the results for secondary endpoints are subject to inflated type 1 error rates.”

Additionally, the reviewers noted that results for the secondary endpoints “all contain an upper bound near the zero point, indicating no effect,” and that this observation that “casts further doubt on the claim of benefit.” In any case, they wrote, the “small size of the point estimates would be of unclear clinical significance.”

Even those who applaud the conversion of accelerated approval of Elevidys to a full approval should be concerned about the process.

There are two ways of looking at the decision by Marks to overrule the adamant objections of CBER’s review staff.

One is that he is right and his staff, who carefully weighed the evidence, came to the wrong conclusion, and that the standards they used to judge the evidence are faulty.

The other is that the staff were correct, and Marks got it wrong.

The first possibility suggests that CBER staff, from top to bottom, misunderstood the data, made an egregious mistake, and that the center’s director believes they can’t be trusted. This should be disturbing to the public because Marks can’t pore over every marketing application to double-check his staff’s work, and because it raises the specter of sponsors going over the heads of review staff and appealing to senior managers, a practice that corrodes morale.

Many companies that had data that were more “compelling” than Sarepta’s have had BLAs rejected. They, and patients, deserve to know what standards CBER will apply in the future.

The second option, that Marks consciously or unconsciously cherry-picked data to support a predetermined outcome is also bad. It would mean that boys with DMD are being exposed to a treatment that presents real safety risks that are not offset by benefits.

Part of the problem is that Marks didn’t describe his disagreement with CBER staff as a matter of policy. Rather, he said that he disagreed with their conclusions that “the data submitted do not confirm the benefit of Elevidys” and that Sarepta “provided no satisfactory data to support effectiveness claims for all ages and for non-ambulatory patients.”

Similarly, Marks dismissed a CBER reviewer’s assertion that the secondary endpoints Sarepta pointed to as evidence of efficacy were not prespecified for hypothesis testing and as a result “cannot be attributed to an actual treatment effect.”

If Marks had articulated his rejection of the review staff’s conclusions as a shift away from rigid adherence to certain statistical policies in favor of regulatory flexibility in cases where the unmet medical need is great and a pattern of modest efficacy is shown across secondary endpoints, that would have been useful information for other companies and patient communities. It would also have been in keeping with recent decisions made by different groups within FDA. 

Instead, he not only missed an opportunity to build trust in the coherence of the agency, but also publicly undermined his staff.

Janet Woodcock has suggested the establishment of a new pathway for approving drugs that cannot be tested using traditional methods. If FDA is going to go down that route, it should be debated and legislated. In any case, the problem with DMD therapies doesn’t seem to be that it isn’t possible to test them using conventional methods, it is that the effects of therapies are modest or non-existent. Sarepta claims, and Marks has accepted, that a change in time to rise from the floor of six-tenths of a second and of four-tenths of a second in the 10-meter walk/run over a 52-week period compared to placebo is sufficient to merit approval.

It may have been possible once upon a time to dismiss a center director’s decision to override agency staff as a one-off event that didn’t create a precedent. Now there have been several instances going back to 2016, when Woodcock, then director of CDER, overruled her staff to approve Sarepta’s Exondys 51 for DMD.

It is relevant to note that the company still hasn’t completed a trial that confirms or refutes the clinical benefit of Exondys 51, and that in the absence of confirmatory evidence FDA has nevertheless deemed micro-dystrophin expression an approvable endpoint.

The micro-dystrophin controversy has been bubbling for over eight years.

The DMD community, drug developers and FDA would benefit from a transparent, science-driven discussion of the issues. A workshop the Reagan-Udall Foundation for the FDA held in February about biomarkers for ultrarare diseases could service as a template.

Controversy over the Elevidys episode puts at risk the progress Marks and CBER are making in advancing cell and gene therapies for rare diseases.

However it goes about it, FDA should get its house in order so its senior management, patients and the public can rely on the decisions made by officials who have dedicated their careers to the science-based regulation of medical products.