New FDA papers reveal internal split over Biogen's Aduhelm approval

The FDA has documented the steps leading up to the approval of Biogen’s Alzheimer’s disease drug Aduhelm, revealing how the mixed evidence on the controversial therapy divided its staff.

In its summary of the Aduhelm review, the FDA shares information about how the divergent opinions among its staff that were apparent at the November advisory committee meeting played out across internal meetings in March and April. The process culminated in an April 26 meeting where senior FDA leaders discussed plans to grant accelerated approval to Biogen in light of doubts about whether it had provided substantial evidence of effectiveness on the clinical endpoints. 

Most of the leaders in attendance supported the approach. Crucially, Patrizia Cavazzoni, M.D., Peter Marks, M.D., Ph.D., and Richard Pazdur, M.D.—respectively, the directors of the FDA’s centers for drug evaluation, biologics evaluation and the oncology center of excellence—voiced support for accelerated approval. Pazdur has experience of accelerated approvals but specializes in cancer, not neuroscience. 

The proposal was opposed by Sylva Collins, Ph.D., director of the Office of Biostatistics, who argued there was insufficient evidence to support accelerated approval or any other type of approval. Shahvree Buckman-Garner, M.D., Ph.D., director of the Office of Translational Sciences, said she understood both sides of the argument. Buckman-Garner’s office covers both biostatistics and clinical pharmacology.

The divisions at the meeting in late April reflect the split between the FDA’s statisticians and the rest of the agency that was seen at the advisory committee meeting months earlier. At the event, Tristan Massie, Ph.D., a mathematical statistician at the FDA, said the “whole pattern of study outcomes could be explained by a randomly worsened placebo after the dose increase amendment,” adding that there is “compellingly conflicting phase 3 evidence,” and approval “would likely present challenges to other Alzheimer’s candidates in development.” Massie recommended against approval.

In the end, the objections of Collins and Massie on biostatistical grounds were overruled. The files released by the FDA describe no internal discussions after April 26, although the agency did meet with Biogen early the next month to discuss the training program for detection and management of amyloid-​related imaging abnormalities. 

The FDA published its summary of the review alongside statements from Cavazzoni and Peter Stein, M.D., director of the FDA’s Office of New Drugs, that set out their support for the accelerated approval. Stein used his statement to challenge Massie’s view that a worsened placebo response can explain the seemingly positive effect of Aduhelm in one trial.

“I find this argument problematic since it compares results from subsets of patients within a treatment group rather than from comparison between intact treatment groups. Further … the effect size on CDR-SB relative to placebo pre- and post-amendment are very similar, not supporting the concern raised by Dr. Massie on this point.”

Stein accepted that giving Aduhelm accelerated approval “veers” from the typical scenario, in which the pathway is used when it would take years to generate evidence against a clinical benefit endpoint to support a traditional authorization. Biogen tried to show clinical benefit, failed to meet the bar for a traditional approval and was bailed out by the FDA’s conclusion that the data satisfy the criteria for accelerated authorization.