So it’s finally time for Biogen to sit down with an FDA advisory committee to look at their proposed Alzheimer’s therapy, the anti-amyloid antibody aducanumab. I last wrote about it here, back in December, and you know what? I haven’t changed my mind a bit, because (1) no new data have emerged (none were expected) and (2) I have not had a change of heart about the existing numbers. You can read the post and the links in it for more, but to put it briefly: I do not think that Biogen has demonstrated efficacy. I think that they have enough of a hint to run a better confirmatory trial, should they so desire, but they do not. They desire to go to the FDA, get the drug approved, and begin printing money.
And I would be all for a drug company printing money if they had a drug that could really alter the course of Alzheimer’s disease, but (once again) Biogen has not, in my opinion, demonstrated that they have anything like that. And I am definitely not all for a drug company printing money for something that really doesn’t do anyone any good. Because everyone knows what’s going to happen if aducanumab is approved: the pent-up demand for something, anything to treat Alzheimer’s is immense. Has been immense forever. There are a lot of people who have a family member with the disease, and they will demand treatment with the new drug that the FDA has approved to fight the disease. Who could blame them?
The FDA has released its briefing documents for the advisory committee meeting, and they took a lot of people by surprise (I was one of them). A key document is here, and a key section of it starts around internally numbered page 147 of the PDF. That the FDA clinical reviewer’s commentary on the efficacy data in the trial, and let’s say that he’s a lot more convinced than I am. Here’s a writeup on it at Stat, from Damian Garde and Adam Feuerstein.
There are two studies under discussion: 301 and 302. Broadly speaking, 301 (also known as ENGAGE) was negative, and 302 (also known as EMERGE) was positive, and the problem has been which result to believe and to account for why they’re different. There’s also a problem in that the data are incomplete; Biogen itself stopped work in the clinic when a futility analysis suggested that the antibody was not working, but revived their hopes with a post hoc analysis. (Here’s my writeup on this at the time). But the FDA clinical reviewer is very persuaded by the positive results, and believes that the negative trial was skewed by a population of fast-deteriorating patients. As I recall, Biogen’s rationale was that the negative trial looked better if you went for the subset of patients who got sufficient exposure; I don’t think they had advanced the fast-deteriorating-cohort explanation at the time, although I’d be glad to be corrected about that. Those rapid progressors seem to have mostly been in the high-dose arm of 301, and to be fair, that is where the greatest problem with the results showed up. And I mean “problem” in the sense of “worse than placebo”. But with the new analysis, the clinical reviewer is unbothered by 301, adds its good parts into 302, and concludes that the data are “robust and exceptionally persuasive”.
This is what goosed Biogen’s stock price the other day, as well it might. But even at the time, people were starting to notice the comments of the FDA’s statistical reviewer, which start on internally numbered page 247 of the same PDF. “Inconsistency on many levels summarizes the final clinical efficacy data from these trials,” is how things start off there, and the review goes on to note (among other things) the missing data at week 78 (due to the trial halt, and the time point that had shown the most evidence of benefit). “It is not justifiable to search 301 for patients who are most similar to 302“, it goes on to say, “because that may have selection bias and gives the impression that 302 is right and 301 is wrong, for which there is no justification not relying on post-hoc analyses“. Statements like that, and there are many, give one the strong impression that this reviewer had read at least a draft of the clinical reviewer’s work. Biogen, the review goes on to note, is advancing that post-hoc “rapid progressor” theory, but the reviewer says that after the completion of the study is too late to try to address that – and besides, an effective drug would be unlikely to fail so completely just due to rapid progressors in what was considered a study in people early in the disease.
Overall, the statistical reviewer says that the negative 301 study “should not be discounted without some extremely compelling reason (which there is not)”. Overall, after working through a number of the issues in the trial, the conclusion is that “. . .the totality of the data does not seem to support the efficacy of the high dose. . .the reviewer believes that there is no compelling substantial evidence of treatment effect or disease slowing, and that another study is needed. . .” The figures that follow in the PDF are worth a look too, if you’re into this sort of thing. They get into a great deal more detail than I have time or space to here in this post, but they constitute a sustained attack on the hypothesis that that aducanumab has shown efficacy and the interpretability of the data presented for that case.
The fact that Biogen stock jumped on the release of this document tells you most of what you need to know about the stock market (and about investing in biotech stocks in particular). You wonder how many people even got as far as the statistical review section before hitting the big green Buy button. Now, I have no idea of the internal politics of this FDA review – and since the place is staffed by humans, it will have politics. But this looks like a deeply divided briefing document – frankly, I can’t recall seeing one that was more at odds with itself. I don’t know what the advisory committee will make of all this, either, although it has to be noted that one of the AdComm members, David Knopman of the Mayo Clinic, has been recused by the FDA and will not sit. He was involved in the Biogen trials, which is indeed a conflict of interest – but he emerged from the experience as a critic of the drug who thinks that another trial is the only way to answer the questions about it.
I’ve been worried about this sort of situation in Alzheimer’s for years, and I’m not the only one. I really wish that we weren’t fighting out a big approval decision under such conditions, but here we are. I have no idea of what’s going to happen in the advisory committee, although the loss of Dr. Knopman increases the chance of a favorable vote. And beyond that, I have even less of an idea of what’s going to happen with the eventual approval decision. I don’t think that the FDA should approve aducanumab as it stands, in the same way that I don’t think that they should approve any drug for which the evidence is this equivocal. But the agency approved Exondys, so who knows what they’re capable of? We’ll find out. Lucky us.
And if they do, we’ll find out how many physicians will prescribe it, and how many insurance companies will pay for it. It would be something to see both of these groups hold the line, but I fear that the pressures will be just too great. Biogen is counting on just that, and I’m not happy about it.