Lilly’s Alzheimer’s Data For Donanemab

Summary:

• The clinical trial was able to demonstrate a slowing of disease progression, but it’s still not possible to say if that reaches a level that any nonspecialist observer would be able to notice outside of a controlled clinical trial.
• The Lilly trial sorted patients according to their tau protein levels on top of the amyloid deposition that is usually used to measure Alzheimer’s, which could well be useful.
• Donanemab seems to have had less effect on the high-tau patients – in fact, the high-tau group did not reach significance in the study’s primary endpoint.

I wrote earlier this year about donanemab, the anti-amyloid antibody developed by Eli Lilly (NYSE:LLY) that they’re hoping to get approved by the FDA. At the time, I said that we all really need to see more clinical data than the Lilly press release had in it at the time, and now that day has arrived. So what’s the verdict?

Unfortunately, these numbers do not get rid of my doubts. The clinical trial was able to demonstrate a slowing of disease progression, but it’s still not possible to say if that reaches a level that any nonspecialist observer would be able to notice outside of a controlled clinical trial. Just to be clear, by “nonspecialist observers”, I mean family members and acquaintances: would they be able to tell that a person was taking donanemab or not?

It’s still not clear, based on the changes in the rating scales, if they would, and that’s using the rating scale difference after 76 weeks of treatment. It’s important to remember that both the patients getting the antibody and the control patients getting placebo IVs deteriorated steadily over this whole period – it’s a painful and horrible fact that we don’t know of anything that keeps Alzheimer’s patients from deteriorating steadily. At best, donanemab might slow this down a bit, but I have to say that this might also depend on which rating scale you use. The ADCAS-iADL scale shows the greatest effect, albeit with the widest variance, but others (such as the MMSE) show what may be less slowing of the disease progression (ranging down to imperceptible levels), although to be fair, the error bars overlap. And no, we don’t know which of these scales is the best marker for real-world effects, nor to what degree deterioration in them is likely to be noticed by non-clinicians.

The Lilly trial sorted patients according to their tau protein levels on top of the amyloid deposition that is usually used to measure Alzheimer’s, which could well be useful. The low/medium tau cohort deteriorated more slowly overall, and indeed the low-tau patients seem to have started out with somewhat better scores on the rating scales. Likewise, donanemab seems to have had less effect on the high-tau patients – in fact, the high-tau group did not reach significance in the study’s primary endpoint. Lilly is apparently saying that that’s no reason not to give them the drug anyway, naturally, and that there’s no need to test for tau levels before putting people on it.

There is (again unfortunately) one way that outside observers might be able to guess that someone was taking donanemab, though, and that’s the side effects. Like the other antibodies of this class, this one has a problem with cerebral edema, as seen via imaging abnormalities, as well as outright microhemorrhages (ARIA-E and ARIA-H, respectively). 2.1% of the placebo group showed ARIA-E, as opposed to 24% of the treatment group, with zero serious adverse events of this kind in the former and 13 in the latter. As for ARIA-H, 13.6% of the placebo group showed such imaging results, and 31.4% of the treatment group did, with zero serious events in the former and four in the latter. Being heterozygous (or especially) homozygous for the APOE gene was associated with both sorts of trouble. Overall, the use of anti-clotting drugs did not seem to be associated with these events. 3.7% of the placebo group discontinued treatment due to adverse events, while 13.1% of the treatment group discontinued.

It was not in Lilly’s slides at the meeting where they just presented, but the JAMA publication (linked in the first paragraph does mention cerebral shrinkage, which has been a constant feature of anti-amyloid treatments. Those two posts will give you some more background, but suffice it to say that this is not an effect that seems to be explained just by the loss of amyloid and that it is hard to imagine how it’s a desirable feature of treatment. You have to go to eFigure 6 in the third supplement file, but it looks like whole-brain volume continually decreases through the whole 76-week period, largely driven by increases in (empty) ventricular volume.

Lest you think that I am uniquely downbeat about the prospects for this drug and the other anti-amyloid antibodies, here’s a quote from the commentary article in JAMA accompanying the Lilly publication:

These results serve to highlight the complexity of Alzheimer disease itself. The exceptional ability of drugs such as donanemab and lecanemab to remove amyloid, paired with their rather subtle effect on the rate of decline in cognitive and functional measures, suggests that amyloid is likely not the only factor that contributes to Alzheimer disease progression.

From a more practical standpoint, the modest benefits would likely not be questioned by patients, clinicians, or payers if amyloid antibodies were low risk, inexpensive, and simple to administer. However, they are none of these. . .Whether the harms of these drugs are balanced by their modest clinical benefits will ultimately require more data. . .

Exactly. We will eventually find out if these drugs can actually help patients in the real world, or whether they do more harm than (any) good. In a more rational world, we might try finding all this out before we approve them and sell them to patients, but we’re clearly going to do it the other way.

Disclosure: None

Original Post

Editor’s Note: The summary bullets for this article were chosen by Seeking Alpha editors.