Has the FDA been approving some drugs without proper evidence?

Balancing safety while also supporting innovation to help patients access new therapies means that the pendulum at the Food and Drug Administration often swings between two extremes: fast tracking new drug applications without a ton of data or slower, data-driven approaches that emphasize patient protection.

Lately, that pendulum seems to have swung toward the former of the two, with the agency sometimes approving drugs on the basis of what are seen as weak data. That has led to a perception that the FDA – for better or worse – is prioritizing speedier market entry of some products over ensuring that available data fully support their risk-benefit profiles.

One industry observer traces it all back to a decision three years ago.

In September 2016, the FDA granted accelerated approval to Sarepta Therapeutics’ Exondys 51 (eteplirsen) in Duchenne muscular dystrophy, or DMD. The decision was made over the objections of an expert panel that found the data didn’t sufficiently demonstrate the drug’s efficacy – and even over the objections of the agency’s own reviewers.

Exondys 51 represents just one of several FDA approvals – including accelerated approvals – in recent years that had the agency going against the objections of external as well as internal experts and authorizing drugs backed by questionable evidence.

“I feel like the beginning of the end of rational drug approval began with [Exondys 51],” said Dr. Adriane Fugh-Berman, a professor of pharmacology and physiology at Georgetown University, in a phone interview. “Our standards for drugs have gone down and down.”

A Phase II study of Exondys 51 had shown an increase in the protein dystrophin, a surrogate endpoint, but not a definitive clinical benefit. That didn’t convince the FDA’s Peripheral and Central Nervous System Drugs Advisory Committee, and it hasn’t convinced the European Medicines Agency either. But FDA’s Center for Drug Evaluation and Research Director Janet Woodcock overruled agency reviewers’ objections and granted accelerated approval anyway. She cited a need for the “greatest flexibility possible” in the face of a fatal disease without approved treatments. But it’s been widely suggested that pressure from the patient community — including bringing several boys in wheelchairs into the AdCom meeting — helped sway the agency.

FDA spokeswoman Sandy Walsh countered the narrative that FDA is expediting approvals without data to back it up.

“Regardless of any expedited pathway or designation, we only approve drugs when the data received in a drug application demonstrate a favorable risk-benefit profile,” Walsh wrote in an email. “Patients – especially those with difficult-to-treat diseases – often have few or no therapeutic options, and we take that into account when examining the risk-benefit profile of those drugs.”

She specifically addressed the Exondys 51 approval. 

“We take the views of the patient community into account when evaluating the benefits and risks of a drug,” Walsh wrote. “The patient testimonies provided the opportunity for the FDA to learn more about patient experiences with the disease. Their personal experiences provided helpful insight into how serious a condition DMD is.”

Officially, the FDA is not bound by AdCom votes, but usually follows them. A study published last month found that it went against those votes about one-fifth of the time, based on data from between 2008 and 2015. Yet, the only predictive variable the authors found was that in those cases, the vote was usually divided.

The same month the study was published, the FDA went against an Oncologic Drugs Advisory Committee vote and gave accelerated approval to Karyopharm Therapeutics’ Xpovio (selinexor) for so-called “penta-refractory” multiple myeloma patients who have exhausted five drugs across three major drug classes used for the blood cancer. That move drew objections from oncologists who argued its safety and efficacy profile did not warrant the approval.

“It is certainly true that we want to accelerate the development of promising therapies for patients with hard-to-treat cancers,” Stanford University hematologist-oncologist Dr. David Iberri wrote in an email, calling Xpovio a “poster child for what’s wrong” with the FDA’s “liberal” approach to oncology drug approvals. “But what’s happening now is that drugs are being approved before they’re fully developed, before their promise has been demonstrated adequately in studies.”

A major factor in the FDA’s reversal on Xpovio was that the Phase III study’s data safety monitoring board submitted confidential data from the study to the agency.

Although it received full rather than accelerated approval, Otsuka Pharmaceutical’s Abilify MyCite – approved in November 2017 – has also drawn criticism over what some see as weak data, even though the FDA initially rejected the application. The drug is a version of the schizophrenia and bipolar disorder drug Abilify (aripiprazole) that incorporates a chip from Proteus Digital Health to monitor whether patients have taken it and is considered a digital medicine.

But in a paper published in July in the British Medical Journal, a team of researchers looked at the data used to support the drug’s approval, concluding that it was based on weak evidence because there was no data indicating it produced better adherence, including the lack of a randomized, controlled trial comparing the digital drug with the regular drug, other active comparators or placebo. Consequently, they raised the question of whether Abilify MyCite’s development was meant to “evergreen” a drug that is available as a generic.

Lead author Lisa Cosgrove, professor of counseling and school psychology at the University of Massachusetts Boston, also cited concern about a schizophrenia drug with a digital sensor that tracks usage.

“It’s a drug used primarily to treat people who are paranoid that people can read their thoughts or have things inside of them,” she said in a phone interview.

However, Otsuka spokesperson Robert Murphy wrote in an email that the study was based on an “inaccurate premise,” stating that the company never requested a claim for “improved adherence” when it sought approval for Abilify MyCite. The Proteus sensor can provide data regarding medication-taking behaviors and other activities, thereby enabling clinicians to make more informed therapeutic decisions and giving patients a greater role in their treatment and well-being. Both the drug and sensor were independently approved by the FDA, and Otsuka submitted studies to assess the use of the system in a representative sample of intended users, including schizophrenia patients, he wrote. It also convened a panel of bioethicists to address privacy and ethical issues concerning those patients.

To be sure, these drugs do not represent the totality of the FDA’s recent approval history. And accelerated approval has enabled many important therapies to get to patients much faster than they otherwise would have. Under accelerated approval, the FDA expedites market authorization for a drug based on data that are early or based on so-called “surrogate” clinical endpoints, which indicate but don’t yet prove a clinical benefit. Accelerated approval is conditional on the product’s sponsor providing that proof, usually in the form of a larger, randomized, controlled trial. Other tools meant to speed along development include the FDA’s Breakthrough Therapy, Fast Track and Priority Review programs.

But the aforementioned drugs have raised questions about whether the FDA has become too lenient in its standards, especially regarding products deemed to address unmet medical needs or seen as representing a significant technological advancement.

Fugh-Berman, the Georgetown pharmacology professor, partly blamed what she called the FDA’s falling standards on the the Prescription Drug User Fee Act of 1992, also known as PDUFA, which charges drugmakers a user fee when they file for approval.

“Having drug companies pay for the approval of their drugs has given them a lot of [leeway] at the FDA, and there has been a loosening of standards for drug approval that is not going to be good for public health,” she said.

While not wanting to commit to whether the FDA has become overly or insufficiently restrictive, Dr. Joseph Ross, a professor of medicine at Yale University, said Congress and patient groups have been advocating over the last 30 years for progressively lower evidentiary thresholds.

“My big takeaway is that I’m cautious – I would like to see better evidence at approval,” he said in a phone interview. “What we should be doing is imposing much stronger post-marketing requirements to generate evidence.”

But Jennifer Miller, an assistant professor at Yale who teaches bioethics, sees it in more cyclical terms. “If you look at the history of drug regulation, there’s somewhat of a pendulum that swings between preferring protecting patients by ensuring that trials are adequately controlled and experimental interventions are rigorously studied, versus the other side of the pendulum, which is speeding innovation in order to ensure access to drugs,” she said in a phone interview.

For example, patient protection was the order of the day in the early 1960s, when severe birth defects overseas from the drug thalidomide led to a tightening of drug regulations in the U.S.

The current regulatory milieu favors getting drugs to market faster. Of course, the FDA still carries the burden of ensuring drugs are safe and effective. However, Miller said, people are willing to assume a higher level of risk and also forego the traditionally paternal relationship between doctors and patients, as patients want a greater role in decision making. One example is the Right-to-Try movement, which scored a big win last year when President Donald Trump signed a federal right-to-try bill into law, after many states had already enacted bills of their own. Under the law, in certain situations patients can obtain drugs from manufacturers without any oversight from the FDA at all.

Greater attention to serious and orphan diseases and increased public concern about reducing drug costs over the last several years are likely factors contributing to a more liberal regulatory regime, said Chad Landmon, who chairs the FDA practice at the law firm Axinn, Veltrop & Harkrider.

“There has been a historical perception that it takes too long to get products approved,” Landmon said, in a phone interview, adding he’s been to conferences where academic experts and industry figures have speculated that getting aspirin approved under the current regulatory system would be virtually impossible.

So he views the tendency at the FDA toward speedier approvals as a net positive.

A review of FDA data last month by The Wall Street Journal indicated that more than 60 percent of drugs approved in the last five years have received accelerated approval; during the previous five years, it was less than 60 percent.

The trade-off to allowing speedier approvals is that it creates a need for more confirmatory clinical trials, including Phase IV studies, Landmon and Miller said.

The problem is, sometimes that doesn’t happen.

While Exondys 51 has been on the market for three years, Sarepta has yet to provide confirmatory data for it.

Walsh responded that the agency’s recent report shows that the majority of post-marketing studies were successfully completed or progressing toward completion according to the original schedule in fiscal year 2018, with very few delayed.

Even when there is data leading to final approval, it doesn’t always rely on “hard” endpoints that definitively prove a benefit. In 2017, the FDA granted full approval to Roche’s Avastin (bevacizumab) for glioblastoma — it had received accelerated approval in 2009 — on the basis of progression-free survival (PFS) data. While the PFS benefit means the cancer has stopped growing, Avastin has never shown a proven benefit on overall survival, the amount of time a patient actually lives – a much more important endpoint than PFS in a disease with a median survival of less than two years from diagnosis.

A study published in May in the Journal of the American Medical Association found that only 19 of 93 cancer drugs receiving accelerated approval over a 25-year period ultimately extended OS, generally considered the gold-standard endpoint in cancer.

No one can say if the current regulatory regime will lead to a major issue imperiling patient safety. After all, the FDA’s standards today are far stronger than they were in the mid-20th century.

Still, a swing of the pendulum toward stronger regulations would likely take nothing short of a disaster, Miller said. She cited a bioethics maxim: “Ethics is born in scandal and reared in protectionism.” High-profile events involving children – such as the thalidomide scandal – are often a trigger for stricter regulations, she said.

Despite that often being the case, it may not take such a scandal to pull back the reins at the FDA. 

On Monday, the agency declined to approve Sarepta’s second DMD drug, Vyondys 53 (golodirsen), based on data from a Phase I/II study of 39 patients. The official reason was concerns over kidney toxicity and injection-site infections, but an investment analyst speculated that it may have been a “slap on the wrist” after the controversy surrounding the approval of Exondys 51.

Photo: Food and Drug Administration, via Flickr (free of all copyright protection)