Newly disclosed FDA documents reveal agency’s unprecedented path to approving Aduhelm

Chastened after a decisively negative review from outside advisers, the Food and Drug Administration convened a series of internal meetings in March and April where top officials hammered out a plan to approve Biogen’s Alzheimer’s drug, Aduhelm.

The meetings were revealed in a series of documents released Tuesday by the FDA to explain its decision to use a truncated pathway, called accelerated approval, to approve Aduhelm.

The document dump follows weeks of bracing criticism of the FDA, which departed from regulatory precedent to approve Biogen’s treatment. Instead of judging Aduhelm based on its effect on the progression of Alzheimer’s, for which the evidence is debatable, the agency approved the drug based on its ability to remove brain plaques called beta-amyloid, which are believed to contribute to the disease. One agency adviser, resigning from his committee post in protest, called it “probably the worst drug approval decision in recent U.S. history.”

“We strongly disagree with that characterization of our decision,” said Patrizia Cavazzoni, director of the FDA’s drug division, in an interview with STAT on Tuesday. “We think the decision is on very solid ground, that we are on very solid ground when it comes to the data and the rationale for utilizing accelerated approval to greenlight this drug.”

Exacerbating the controversy was Biogen’s decision to set a $56,000-a-year list price for Aduhelm. More than 1 million Americans are likely eligible for the treatment, making Biogen’s product a potential budget-buster for Medicare.

The FDA’s controversial decision was the product of months of internal deliberation, according to 83 pages of internal agency documents released Tuesday.

After the November 2020 meeting in which its outside advisers recommended against approving Aduhelm, the FDA spent months locked in debate over the evidence supporting Biogen’s treatment. To one contingent, Biogen’s data — culled from two clinical trials, one positive, the other negative — were too messy to support the notion that Aduhelm delayed the progression of Alzheimer’s. To another, the evidence, combined with a desperate desire for new treatments, was enough to justify approving the drug.

In two meetings, on March 31 and April 7, FDA officials presented Biogen’s application to the Medical Policy and Program Review Council, an internal group that helps set agency policy. According to an FDA summary of the council meeting, what emerged was a third path: While it’s not clear that Aduhelm actually slows down Alzheimer’s, there was “convincing evidence” that it had a “robust” effect on amyloid plaques.

The council concluded, based on the current understanding of the disease, it’s “reasonably likely” Aduhelm’s effect on amyloid is predictive of a cognitive benefit, which “supports the accelerated approval of” Biogen’s drug, according to the FDA’s summary.

This would break agency precedent. Accelerated approval is traditionally used for treatments that haven’t yet proved themselves in large trials. In Biogen’s case, Aduhelm went through two Phase 3 studies and came up with conflicting evidence. But the council, citing federal law, concluded that the FDA had the authority to use accelerated approval on any treatment “upon a determination that the product has an effect on a surrogate endpoint that is reasonably likely to predict clinical benefit.” Amyloid fit the bill.

“We recognized this actually fit very well into that accelerated approval paradigm,” said Peter Stein, director of the FDA’s Office of New Drugs, in a Tuesday interview with STAT. “This was not us trying to fit a round peg into a square hole. This is a program that fit very well.”

With that idea in mind, on April 26, seven top FDA officials from across the agency met to review granting an accelerated approval to Aduhelm. Cavazzoni and Peter Marks, head of the FDA’s biologics division, were in favor. So too was top FDA oncologist Richard Pazdur, who was brought in because of his experience using accelerated approval for new cancer treatments. Directors in charge of the FDA’s clinical pharmacology and medical policy divisions also agreed.

The leader of the agency’s office of translational sciences abstained. Sylva Collins, director of the FDA’s office of biostatistics, was the only dissenting voice.

The agency’s summary describes no further meetings on the topic. Six weeks later, on June 7, the FDA announced the accelerated approval of Aduhelm.

Absent, but referenced in other memos, were arguments from the FDA statisticians that the drug should not be approved. Also included in the documents are references to another analysis of six other Alzheimer’s drugs that the FDA helped use to justify the decision. Cavazzoni said that the FDA is releasing some documents ahead of its usual timelines because of the intensity of the public debate, and that more detailed information about the statistical review and the review of other Alzheimer’s drugs will follow.

Stein wrote that “the evidence is not sufficiently compelling or persuasive to meet the substantial evidence standard for standard approval.” However, he noted later: “Nonetheless, I agree that the results discussed above strongly suggest that treatment with aducanumab may result in clinical benefit.”

FDA officials acknowledged the accelerated approval is unusual because it is usually used when studies assessing efficacy are not completed. In the case of Aduhelm, the issue is that one of these studies had failed.

Stein wrote that the uncertainty about the benefits of a drug approved via the pathway are “typically because the studies assessing clinical benefit are ongoing, but here because the results of the studies assessing clinical benefit strongly suggested but did not establish benefit.”

Another document, from the FDA’s office of neuroscience, also appeared to wrestle with the implications of the decision. Would approving the drug mean that companies would attempt to salvage failed trials by pointing to a so-called surrogate endpoint after-the-fact? Would this result in less effective drugs?

But the FDA officials conclude that “the circumstances here are fundamentally different,” arguing that there is evidence to suggest that removal of amyloid plaque will predict a clinical benefit. The officials argue that there are data to suggest a clinical benefit of the drug, and that amyloid is linked to Alzheimer’s.

“This is not the typical run-of-the-mill situation,” Stein said. “But just because it’s not the typical situation doesn’t mean that the fit was not quite, quite, quite good for what accelerated approval is built for and how it’s to be done.”

Cavazzoni reiterated that the agency is “very confident that the data are solid.” She added: “In fact, because the size of this program was so large, including a little bit over 3,000 patients in the safety database, this is a situation where the usual uncertainty that we would have in an accelerated approval situation is actually less because we have so much clinical data.”