Of 853 people given Lilly’s Alzheimer’s drug donanemab in the trial behind its approval, 307 showed brain abnormalities on their scans. That is 36 percent. Among 874 people on placebo, 122 did, or 14 percent.
The drug is sold as Kisunla, and the US Food and Drug Administration cleared it in July 2024 for early-stage patients, meaning mild cognitive impairment and mild dementia. The studies behind the decision enrolled people aged 59 to 86 whose PET scans confirmed amyloid in the brain. It is not meant for later-stage disease.
What the scans showed
The abnormalities have a name, ARIA, short for amyloid-related imaging abnormalities. They come in two kinds. One is swelling of brain tissue, found in 24 percent of treated patients against 2 percent on placebo. The other is traces of bleeding, found in 31 percent against 13 percent.
Most cases cause no symptoms and are seen only on an MRI. About 6 percent of treated patients, 52 people, did feel something, and in roughly 85 percent of those (44 of 52) the symptoms cleared. Larger bleeds, over 1 centimetre across, appeared in 4 treated patients (0.5 percent) and 2 on placebo (0.2 percent). The account in the FDA label also notes that this class of drug can produce swelling serious enough to mimic a stroke, and that carriers of two copies of the ApoE ε4 gene run a higher risk. Doctors are advised to test for it before treatment starts.
The pivotal trial recorded three ARIA-related deaths among donanemab patients and none among placebo patients, a figure Tech Times drew from the prescribing information.
A slower start to dosing
Lilly has since changed how treatment begins. Presenting at the 2026 Alzheimer’s Association International Conference in London, the company reported results from a study of 843 participants. Under the original label, patients held at 700 mg through the opening three infusions. The modified schedule climbs from 350 mg to 700 mg to 1,050 mg before reaching 1,400 mg.
At 24 weeks, swelling occurred in 13.7 percent of patients on the gradual schedule and 23.7 percent on the standard one. At 76 weeks the figures were 15.6 percent and 24.2 percent. Symptomatic swelling was seen in 2.8 percent against 4.8 percent. Amyloid reduction barely differed: 67 percent against 69 percent.
How much benefit?
That is the harder question. A physician review of the approval data describes the slowing of decline in the earlier trial as real but small, and cites researchers who judged it short of clinical importance. The same review says treated patients are projected to lose brain volume toward the level typical of Alzheimer dementia about 8 months sooner than untreated ones.
A Cochrane analysis of 17 studies with 20,342 participants, discussed by The People’s Pharmacy, reached a cooler verdict still. Antibodies of this kind, it found, probably make little to no difference to memory and thinking at 18 months, and they raise the risk of ARIA.
Infusions come monthly, and dosing can be stopped once PET scans show plaque has fallen low enough. By weeks 24, 52 and 76, 17 percent, 42 percent and 60 percent of treated participants met that mark.
Sources
- Donanemab (KISUNLA) FDA-approved for Alzheimer's: Should we recommend it?
- Alzheimer's Drug Kisunla Cuts ARIA Risk Nearly in Half, Two-Year Benefit Holds
- Anti-Amyloid Drugs: Breakthrough or Boondoggle Against Alzheimer's?
