Last month I wrote about the newly approved monoclonal antibody lecanemab for Alzheimer's disease. Lecanemab has killed about 1 in every 592 (3 out of 1777) who have received it. It causes serious brain damage in > 1 in 100 and macrohemorrhage in > 1 in 200. Disturbingly, Lecanemab was approved and determined to be safe and effective by the FDA. And now here comes donanemab seeking approval after killing 1 in 284 and permanently brain damaging 1 in 122.

Donanemab in Early Symptomatic Alzheimer's Disease: The TRAILBLAZER-ALZ 2 Randomized Clinical Trial was published on July 17th. Donanemab is a monoclonal antibody designed to remove amyloid plaques. Amyloid depositing in the brain is thought to be an early event in Alzheimer's disease that triggers an "amyloid cascade" leading to progressive neurological damage. It was hoped that removing amyloid could reverse or at least halt the progression of Alzheimer's disease.

Donanemab was given by intravenous infusion once every four weeks for 72 weeks. Patients received at least two positive electron tomography (PET) scans and at least five magnetic resonance imaging (MRI) studies. So administering this drug and carefully monitoring clients is fantastically cumbersome, impractical, and expensive.

Donanemab both succeeded and failed. It largely succeeded in clearing amyloid. 76.4% of the patients treated achieved "amyloid clearance" by 76 weeks vs only 0.3% in the placebo group. That's an amazing biologic and intellectual feat. Unfortunately, like lecanemab, the removal of amyloid did not cause neurological improvement, did not halt the progression of Alzheimer's, and did not even slow the descent a clinically meaningful amount.

The donanemab group declined 7%, the placebo group declined 9%, so the absolute benefit was a 2% slowed rate of decline. Yet, because 7 is 23% less than 9, it can be said the donanemab slowed clinical progression a relative 23%. That number has been reported out of context to give a false impression of efficacy. Similarly, in one subgroup 32% of the donanemab patients didn't decline over the 72 weeks! But neither did 24% of the control group.

On average, the 622 of out of the original 860 randomized to donanemab who completed the trial scored about 3 points higher out of 144 vs placebo on a test of neurological function called the iASDR. It's important to note that the final analysis was only performed on the 622 who weathered treatment, so even that tiny purported efficacy is a product of a significant attrition/survivorship bias because those that died, withdrew, or suffered serious adverse effects aren't included in the analysis. That is not clinically meaningful efficacy, yet it has triggered a powerful wave of false hope driven by articles in the lay press espousing a "turning point" in Alzheimer's caused by these "breakthrough" drugs.

Importantly, there was an overall inverse relationship between amyloid clearance and neurological function. In other words, despite clearing amyloid in 76% neurological status did not improve and neurologic status continued to decline. About half of the clients who achieved amyloid clearance didn't meaningfully benefit from the drug. Those are important noncorrelations that to me argue against the amyloid cascade hypothesis and reveal curing Alzheimer's requires thinking beyond and outside of amyloid-removing therapies just as everyone is going to be going all in on amyloid-removing therapies.

It is critical to note that amyloid deposits are a common finding in old age and their presence does not portend inevitable neurologic decline. About 28% of people with normal cognition have amyloid deposits visible on a PET scan at age 75 and many remain neurologically intact for years as shown in the image below from Prevalence and Outcomes of Amyloid Positivity Among Persons Without Dementia in a Longitudinal, Population-Based Setting

This is critically important because the Alzheimer's Association is seeking to change the definition from a clinical one based on symptoms to a biologic definition. In other words, they want to scan asymptomatic people and diagnose them with a progressive neurodegenerative disease based on nothing more than a nonspecific imaging abnormality and give them a drug like donanemab that has a 1.5% chance of either killing the patient, causing permanent brain damage, or triggering an anaphylactic reaction. And their justification for doing so, the purported benefit is that those who survived treatment declined 2% slower than placebo after 17 IV infusions, 2 PET scans, and 5 MRIs! I wish I were joking. I also wish I could say that neurologists see the inherent danger and absurdity here, but many of them don't.

Direct Harms

Donanemab directly killed 3 out of 853 (1 in 284) people who got it, permanently brain damaged 7 out of the 853 (1 in 122) who got it, caused anaphylaxis in 3 of the 854, caused symptomatic brain swelling in 6.1%, and serious brain damage (termed serious ARIA) in 1.6%. The paper also completely skipped over the condition and the prognosis of the seven people who suffered serious ARIA whose symptoms didn't improve. These were people with only mild cognitive impairment with years to go until severe dementia.