Experts and industry watchers fully expect the FDA to approve Biogen’s controversial Alzheimer’s disease drug aducanumab—but not because it’s a game changer. In the words of one analyst, aducanumab won’t get the agency’s nod because its data are strong, but rather “in spite of the quality” of those data.
But maybe it doesn’t need to be perfect, says Jim Kupiec, M.D., who previously led clinical research at Pfizer’s neuroscience unit in Cambridge, Massachusetts. Kupiec thinks aducanumab’s approval would pave the way for a group of newer, better drugs for Alzheimer’s, including some he’s working on at ProMIS Neurosciences as its chief medical officer.
“This has a modest benefit. If approved, it will be first in class, but it will never be the best in class,” Kupiec said of aducanumab.
“There is a whole generation of assets that are targeting amyloid in some fashion—antibodies or small molecules—that I’m very confident in saying will provide increased effectiveness and increased safety,” he added. “An approval here will, in fact, provide an environment in which all these additional next-generation anti-amyloid therapies are worked on in a very aggressive fashion with investment dollars providing support for additional phase 3 work.”
A program resurrected
The new drug application for aducanumab will hinge on data from three studies, with an emphasis on a small phase 1b study dubbed PRIME or study 103, and a phase 3 study, EMERGE, also called study 302, that enrolled more than 1,600 patients. The third study, ENGAGE, or study 301, was of “essentially identical design” as EMERGE but was deemed a failure.
“Study 301 is a negative study and does not contribute to the evidence of effectiveness of aducanumab,” but it “may be understood well enough … to not represent evidence that the drug is ineffective,” FDA staff wrote in briefing documents released Wednesday.
Though the ENGAGE study is not part of the package to prove aducanumab’s efficacy, the FDA will ask a panel of outside experts when it convenes Friday how the data affect their consideration of the EMERGE data and whether the EMERGE data can stand on their own as evidence that aducanumab works.
Biogen canned the phase 3 program in March of last year because it failed a futility analysis. Six months later, the company surprised everyone, Kupiec included, when it said it had made a mistake and that it would file aducanumab for approval.
What happened? Biogen looked at the phase 3 data that came in after an independent data monitoring board had started the futility analysis and decided the futility analysis was “incorrect.” The company concluded after a post hoc analysis that the EMERGE trial had actually succeeded, with patients on the highest dose of aducanumab logging a statistically significant improvement on a clinical dementia scale. But it also found that patients in the ENGAGE study did worse than patients taking placebo on that same measure, as well as on a test of cognitive function.
Although futility analyses are routine in drug development, this kind of flip-flop is rare.
“I’ve been in the business for 30 years and I’ve never seen anything like this. I’ve never seen a program in which a futility analysis was called and a program was stopped that a subsequent analysis showed to be positive, so this is highly unusual,” Kupiec said.
From futile to FDA
On Friday, the FDA will also ask the Peripheral and Central Nervous System Drugs Advisory Committee to consider the EMERGE data “independently and without regard for” the ENGAGE study, according to a set of draft questions (PDF) published Wednesday. But that ask raised red flags for some industry watchers.
In a prerecorded presentation to the panel, Tristan Massie, Ph.D., an FDA statistician, raised concerns about ignoring the ENGAGE study while considering the “essentially identical” EMERGE study.
“Study 301 fell dramatically short of study 302, and provides a stark contrast and challenge to the 302 results,” Massie said, according to a transcript of the presentation (PDF). That discrepancy could lead to questions about the reproducibility of experiments, Massie said, adding, “we do not have a single strong study in isolation. On the contrary, we actually have a second trial in which the purported effective dose was in the wrong direction compared to placebo, i.e., numerically worse than placebo.”
“Furthermore, if we select only the better study our estimate is very likely biased, and we already know not consistently repeatable in our experience. Thus excluding data from a large trial without sufficient justification is unscientific, statistically inappropriate and misleading,” Massie said.
Like Kupiec, Massie thinks an aducanumab approval would affect other Alzheimer’s treatments in development—but that it will cause more problems than it solves.
“For example, noninferiority would be questionable against these mixed divergent phase 3 results,” he said. “Outside of noninferiority, the need to demonstrate an add-on effect to this drug could create possibly powering and or practical issues.”
Finally, an approval could make it harder to recruit and retain patients for studies testing treatments that “potentially could have more consistent effects,” Massie said.