An FDA advisory committee voted overwhelmingly against the data presented about the controversial Alzheimer's drug candidate aducanumab Friday.
The agency's Peripheral and Central Nervous System (PCNS) Drugs Advisory Committee said the positive results seen in one of two identical phase III studies of the Biogen anti-amyloid drug cannot be considered alone, but must be taken together with its twin that was clearly negative.
Therefore the positive trial isn't evidence of aducanumab's effectiveness, even when adding data from an earlier phase Ib study to the equation, the committee said in a 10-0 vote, with one member abstaining.
"This analysis seems to be subject to the Texas sharpshooter fallacy," where someone first fires a shot at a barn then paints a target around the bullet hole, said panelist and biostatistician Scott Emerson, MD, PhD, of the University of Washington in Seattle.
The PCNS committee considered data from three studies: two identical phase III trials, (the one with negative findings, called Study 301 at the meeting) and (showing positive results, called Study 302), plus a phase Ib safety and tolerability study known as Study 103 that was offered as supportive evidence of the aducanumab's effectiveness.
ENGAGE and EMERGE were terminated in March 2019 when a futility analysis determined aducanumab was unlikely to outperform placebo at completion. In October 2019, the drug's developer, Biogen, reversed its position, saying a review of previously unavailable data showed the drug actually reduced cognitive decline in EMERGE, but not in ENGAGE.
On Friday, the committee voted on four questions about evidence of aducanumab's efficacy.
- Whether EMERGE (Study 302), viewed independently and without regard for ENGAGE (Study 301), provided strong evidence supporting aducanumab's effectiveness as an Alzheimer's treatment: Eight panelists voted no, one voted yes, and two were uncertain. "There are at least a dozen red threads that raise concern about the data," said panelist G. Caleb Alexander, MD, MS, of the Johns Hopkins School of Public Health.
- Whether the phase Ib trial (Study 103) provided supportive evidence of the drug's effectiveness: Seven voted no, no one voted yes, and four were uncertain, with Emerson pointing out that "every phase II study is so impossibly biased in its treatment effect."
- Whether Biogen presented strong evidence of a pharmacodynamic effect on Alzheimer's pathophysiology: Six panelists were uncertain, five voted yes, and no one voted no. Many of the "yes" voters agreed there was an effect on amyloid pathology, but some pointed out that data about other biomarkers, including tau, were "murky."
- Whether exploratory analyses of the phase III trials, plus phase Ib data and the drug's pharmacodynamic effects, make it reasonable to consider EMERGE data as primary evidence of aducanumab's effectiveness: Ten panelists voted no, no one voted yes, and one was uncertain.
PCNS panelists voiced strong concerns about the framing of the FDA's presentation and the questions they were presented to vote on, especially whether it was possible to view the outcome of one of two identical studies in isolation."I was very, very disturbed" by the FDA's interpretation of the data, Emerson said.
Others pointed out that the wording of some questions limited their ability to voice their views about the evidence that was presented.
In the meeting, Biogen's senior vice president Samantha Budd Haeberlein, PhD, attempted to account for the differences between EMERGE and ENGAGE, suggesting that failure in one trial did not detract from the findings of the other. To a large extent, Billy Dunn, MD, director of the office of neuroscience in the FDA's Center for Drug Evaluation and Research, agreed, echoing what was presented in the FDA's briefing documents released on Wednesday and emphasizing the "urgent" need for Alzheimer's treatment.
But throughout the meeting, the advisory committee members reiterated inconsistencies crystallized by FDA statistician Tristan Massie, PhD: "It's not like we have one strong study in isolation," Massie wrote in an appendix to FDA briefing documents. "We have an equally-sized and identically-designed study, 301, that directly contradicts 302. If you have two [studies] and you take the best and pretend like it's the only one, your estimate is likely biased as in publication bias."
The Friday votes are nonbinding and will serve as recommendations to the FDA, which is scheduled to by March 7, 2021. If approved, aducanumab will be the first novel medication for Alzheimer's disease since 2004 and the first disease-modifying drug for the condition ever.