The FDA detailed its biggest concerns about Eli Lilly’s experimental Alzheimer’s therapy donanemab ahead of a last-minute advisory committee meeting to convene Monday about the drug.
In documents posted online Thursday, regulators offer a generally positive outlook but say they disagree with Lilly’s decision to emphasize its own assessment of cognition and function over a standard measurement, raise questions about the proper time to stop treating patients, note a lower efficacy in patients with more advanced disease, and probe an apparent higher rate of deaths in those who received the drug compared to a placebo infusion.
An additional analysis of deaths requested by the FDA partly quelled those concerns, but brain bleeding and swelling remain real and potentially lethal risks for donanemab, the documents show. Lilly previously reported that three people who showed signals of brain bleeding and swelling, known as ARIA, died in its Phase 3 study.
In its own FDA briefing documents shared with Endpoints News, Lilly detailed the deaths of two more patients that died with ARIA and brain hemorrhages in the long-term extension phase of the study, but argued that there was no “excess death related to donanemab.”
Weighing the drug’s risks and benefits is sure to be a focus of discussion when the outside experts in the Peripheral and Central Nervous System Drug Advisory Committee meet on Monday. It will be an important moment for the Indianapolis drugmaker, which has been trying to make a medicine for the memory-robbing disease for more than 30 years.
Lilly has faced many failures along the way. Donanemab, its most promising drug yet, has been dismissed by the FDA once for a premature application, and the agency’s decision on its full data has been twice delayed. In March, as Lilly was anticipating an approval, the FDA called for an eleventh-hour advisory committee meeting to discuss the drug.
The decision surprised many experts who expected donanemab would breeze through review since the FDA has already cleared two similar antibody therapies that target amyloid proteins that accumulate in the brains of Alzheimer’s patients.
Safety concerns
Observations on brain scans called amyloid-related imaging abnormalities, or ARIA, are the most common and potentially serious side effects of the drug. ARIA-E, indicative of swelling, was found in 24% of patients on donanemab, and ARIA-H, indicative of bleeding, was observed in 31% of patients.
Those rates are about twice as high as seen in Eisai and Biogen’s study of their commercial drug Leqembi. Serious cases, which could require hospitalization, occurred in 2% of treated patients in Lilly’s trial.
Data initially suggesting that deaths may have been more common among people who got donanemab also appear to concern the agency in its briefing. The FDA notes that 17 people on donanemab died compared to 10 who got a placebo. A closer look at the mortality rates revealed that nearly a quarter of participants in the study quit before it was over, and that Lilly didn’t know if they were alive or dead.
The FDA asked the company to follow up. Working through a vendor, Lilly tracked down the status of about half of those patients, but still lacked information on the cause of death. The updated data showed that 16 placebo recipients and 19 donanemab recipients had died, clearing the apparent imbalance apart from the three known ARIA-related deaths.
Briefing documents from both the FDA and Lilly note two additional deaths from a larger pool of safety data that included patients in long-term follow-up studies. Both deaths occurred in people that appeared to have cerebral amyloid angiopathy, a risk factor for ARIA and brain bleeding. And one of the patients who appeared to have a stroke was administered a thrombolytic therapy — a scenario that echoed deaths following treatment with Leqembi.
The FDA also notes that patients who received donanemab had their brains shrink slightly more than those who got a placebo. Although brain volume loss is expected in Alzheimer’s patients, the agency says the significance of this finding is unclear, and that longer-term studies in more patients will be important.
Efficacy disagreements
The FDA says in its briefing that during the course of the Phase 3 trial, Lilly changed the study’s primary endpoint from the standard assessment of cognition and function, known as CDR-SB, to its own measurement dubbed iADRS.
Lilly has previously argued that its assessment, which combines two other cognitive tests, provides a better read on how patients are doing. But the FDA disagreed with the approach and told Lilly not to use iADRS as its primary endpoint. Lilly ignored the advice, but still used the standard scale as a secondary endpoint.
Ironically, the drug slowed the disease by 22% on Lilly’s scale and 29% on the standard scale — in line with the 27% slowing reported for Leqembi.
Despite the similarity, Lilly’s drug was quickly hailed as the most effective yet when it announced results last summer. That claim was largely dependent on a subset of the study, the two-thirds of patients with low-to-medium levels of tau proteins in their brains. In those patients, who are earlier in the course of their disease, the drug slowed decline by 35% on Lilly’s scale and 36% on the standard scale.
But patients who had higher levels of tau, and presumably more advanced disease, saw smaller benefits, with a 21% slowing on the standard scale. The effects were statistically insignificant on three other scales, including a mere 6% slowing on Lilly’s own.
The FDA says that “the potential for benefit may be more limited” in patients with high tau, and says that it will ask the advisory committee to consider if “there are subgroups of patients for whom the benefit-risk assessment may be more or less favorable?”
Lilly opposes restricting its drug, and argued that imaging tau in the clinic would be onerous and impractical. The company said there is no reason to believe that the drug would stop being helpful at certain tau levels, which runs contrary to the opinion of some experts. But the FDA and Lilly agreed that there were no safety discrepancies between patients with higher or lower tau.
Dosing questions
In addition to the rates of ARIA, the frequency of dosing has emerged as another potential differentiator between donanemab and Leqembi.
Unlike Biogen and Eisai, which say that Leqembi should be dosed indefinitely, Lilly doesn’t think patients should continue getting the drug once most of the amyloid plaques have been cleared from the brain.
When patients stopped taking donanemab in clinical trials, reaccumulation of amyloid was slow and there was a continued widening of outcomes in treated versus placebo groups, suggesting that benefits persisted once the drug had done its job, Lilly said in its briefing document.
Lilly also noted that caregivers, doctors, patients and payers have all voiced support for its limited-duration approach to treatment, and wrote that “once the target is cleared from the brain, continued dosing of donanemab is likely not beneficial and only adds to treatment burden and potential risks.”
But the FDA wished that Lilly had done more to prove this, such as including a group of patients in the study who continued to receive donanemab indefinitely. The agency also noted “uncertainty” about how much amyloid should be removed before cutting off treatment.