Lecanemab and donanemab: do they work, and what are the risks?

Who they are for

Both are antibodies given by drip into a vein. They were tested in people with early Alzheimer's (mild cognitive impairment or mild dementia) whose tests confirmed amyloid, the protein that builds up in the brain in Alzheimer's. In the lecanemab trial the drug was given every 2 weeks. In the donanemab trial it was given every 4 weeks, and people were switched to placebo once they met the trial's criteria for completing the dose. [1, 2]

What the main trials found

The lecanemab trial (Clarity AD) followed 1,795 people for 18 months. On an 18-point scale of dementia severity, where higher is worse, scores rose by 1.21 points on lecanemab and by 1.66 points on placebo: a difference of 0.45 points. The researchers called this moderately less decline, and noted that the drug was associated with adverse events. [1]

The donanemab trial (TRAILBLAZER-ALZ 2) enrolled 1,736 people in an 18-month trial. On a 144-point scale where lower is worse, people with low or medium levels of tau lost 6.02 points on donanemab and 9.27 on placebo: a difference of 3.25 points. 23 of its 24 pre-set outcomes were statistically significant. [2]

Is the difference big enough to notice?

This is where the research disagrees. A 2024 review pooled 19 publications with 23,202 participants across 8 anti-amyloid antibodies. It found small improvements on memory and function scales, but none of them, including those for lecanemab and donanemab, reached the smallest change usually counted as clinically important. [3]

A 2026 Cochrane review of 17 studies with 20,342 participants concluded that at 18 months these antibodies probably make little to no difference to memory and thinking (moderate certainty), and may make little to no difference to dementia severity (low certainty). [4]

An open-label extension of the lecanemab trial reported continued benefit up to 36 months. After the first 18 months everyone in it received lecanemab, so there was no placebo group to compare against. [5]

The risks

The main risk is called ARIA: swelling (ARIA-E) or small bleeds (ARIA-H) in the brain, seen on brain scans. The Cochrane review estimated 107 more people with ARIA swelling for every 1,000 treated, and 29 more per 1,000 with swelling that caused symptoms. [4, 6]

The 2024 review put it another way: for every 9 people treated, one extra person had ARIA swelling, and for every 86 treated, one extra person had swelling with symptoms. [3]

In an analysis of 3,030 people in the donanemab trials, ARIA swelling was more likely in people with more copies of the APOE ε4 gene variant, more tiny bleeds already visible on their scans, more amyloid, and higher blood pressure. [7]

Outside the trials, one specialty memory clinic treated 234 people with lecanemab. 23 of them (9.8%) stopped treatment, 10 (4.3%) because of ARIA. The clinic described the frequency of significant side effects as manageable. [6]