Published June 2, 2026 Updated August 26, 2026 10 min read

# Anti-amyloid drugs (lecanemab, donanemab) and what they mean for carriers

## The short version

These drugs can slow early Alzheimer's a little. They are not prevention, and they are not for people without symptoms. Two copies of APOE4 raise the main side effect, brain swelling on MRI, enough that the decision changes with your genotype.

By the OutliveAPOE4 editorial team. [How we research & source](/methodology).

For the first time after decades of dead ends, there are Alzheimer’s drugs that actually slow the disease. The catch lands squarely on you: if you carry two copies of APOE4, the rate of the main side effect (brain swelling seen with magnetic resonance imaging, or MRI) runs about 33% in the lecanemab trial, versus about 5% in non-carriers. So this is a real treatment option paired with a real, genotype-specific risk, and it is one of the few places where your APOE status directly changes a medical decision.

## What this page answers

-   How much lecanemab and donanemab slow decline
-   Why APOE4 changes the risk of brain swelling or bleeding
-   What monitoring and treatment questions to discuss with a specialist

## What these drugs are and how they work

Lecanemab (brand name Leqembi, infused every 2 weeks) and donanemab (Kisunla, infused monthly) are antibodies given by IV. Picture [amyloid-beta](/topics/how-apoe4-affects-the-brain) as a sticky protein that clumps into plaques in the Alzheimer’s brain. These antibodies grab onto that amyloid and flag it so the immune system hauls it away. In large 18-month trials ([CLARITY AD](https://pubmed.ncbi.nlm.nih.gov/36449413/) for lecanemab and [TRAILBLAZER-ALZ 2](https://pubmed.ncbi.nlm.nih.gov/37459141/) for donanemab), they removed a substantial amount of amyloid and modestly slowed the rate of cognitive decline in people with early, biomarker-confirmed Alzheimer’s.

Two things to be clear about. These are treatments for early disease ([mild cognitive impairment](/topics/mild-cognitive-impairment-and-early-detection) or mild dementia with confirmed amyloid), not a cure and not a preventive. An APOE4 carrier with no cognitive symptoms is not a candidate today.

## The benefit, in perspective

The slowing is real, but you should size it honestly. In CLARITY AD, lecanemab slowed decline on the main 18-point clinical scale, the Clinical Dementia Rating-Sum of Boxes (CDR-SB), by 0.45 points over 18 months, a roughly 27% relative slowing. Higher scores on that scale mean greater impairment. Donanemab in TRAILBLAZER-ALZ 2 slowed decline by about 35% on its primary scale. In plain terms, that is a difference of a few months’ worth of progression over a year and a half, not a halt and not a reversal. Whether a patient or family would actually notice that in daily life is genuinely debated among clinicians, and it has to be weighed against the burden of regular infusions and repeated MRIs, plus the risks below.

## The APOE4 catch: ARIA

ARIA is short for amyloid-related imaging abnormalities. It can appear as brain swelling (ARIA-E) or tiny bleeds (ARIA-H) on an MRI.

The side effect that makes your genotype matter is ARIA. As the antibodies strip amyloid out of blood-vessel walls, those vessels can leak. On an MRI that shows up as brain swelling (ARIA-E) or tiny bleeds (ARIA-H). Most of the time there are no symptoms and it is caught only on a monitoring scan, but it can occasionally be serious. A small number of deaths occurred in the trials and after approval, mostly in people who were also on blood thinners, a combination that sharply raises the danger and is often a reason not to mix the two.

Here is where genotype drives the decision, and it does so dose by dose: the more ε4 copies you carry, the higher your ARIA risk, so two-copy homozygotes sit at the top. The gradient is steep. Most of the time these abnormalities cause no symptoms and show up only on a scan, but the difference between one genotype and the next is large enough to change the conversation.

### ARIA risk by genotype, in numbers

In CLARITY AD (lecanemab), the rate of ARIA-E (the swelling form) on MRI rose with each ε4 copy, and symptomatic ARIA-E, the kind you would actually feel, followed the same gradient. Donanemab in TRAILBLAZER-ALZ 2 ran higher across the board, with roughly 40% of its ε4/ε4 homozygotes developing ARIA-E. The figures below come from the [NCBI genetics summary on lecanemab and APOE](https://www.ncbi.nlm.nih.gov/books/NBK605938/) and the donanemab trial program.

**Table:** ARIA-E rates by APOE genotype in anti-amyloid drug trials

| Genotype | Lecanemab: ARIA-E | Lecanemab: symptomatic ARIA-E | Donanemab: ARIA-E |
| --- | --- | --- | --- |
| Non-carriers | ~5% | ~1.4% | lower |
| Heterozygotes (one ε4 copy) | ~11% | ~1.7% | intermediate |
| Homozygotes (two ε4 copies) | ~33% | ~9.2% | ~40% |

So for a homozygote, about a third will show swelling on a monitoring scan and roughly 1 in 9 will have symptomatic swelling on lecanemab, a more honest framing than the trial-wide averages. That is the number worth putting on the table.

### The titration that lowers it

The dose schedule turns out to matter, which is genuinely good news for higher-risk carriers. In TRAILBLAZER-ALZ 6, a phase 3b trial, a slower, stepwise donanemab titration (ramping up over the first few infusions instead of starting near full dose) cut ARIA-E substantially. Overall ARIA-E at 24 weeks fell from 23.7% on standard dosing to 13.7% with the modified schedule. The biggest gain landed exactly where it is needed most: among ε4/ε4 homozygotes, ARIA-E dropped from 57% on standard dosing to about 19 to 24% with the slower titration, roughly a two-thirds relative reduction, while amyloid clearance was preserved ([TRAILBLAZER-ALZ 6](https://pmc.ncbi.nlm.nih.gov/articles/PMC12413699/)). The modified titration is now part of donanemab’s labeling, so it is a concrete option to ask about, not a hypothetical.

### Why carriers get ARIA: the CAA mechanism

The reason ε4 raises ARIA risk traces back to where the amyloid sits. APOE4 is the single biggest genetic driver of cerebral amyloid angiopathy (CAA), the buildup of amyloid in the walls of small brain blood vessels. ε4’s APOE protein binds amyloid-beta about 20-fold more weakly than the ε3 form, so the brain clears it less well and more of it precipitates in vessel walls ([CAA and APOE review](https://pmc.ncbi.nlm.nih.gov/articles/PMC11703089/)). Picture those vessel walls as already partly clogged with amyloid mortar. When an anti-amyloid antibody dissolves that mortar, the vessels become leaky and fragile for a time, which is precisely what shows up as ARIA-E (swelling) or ARIA-H (tiny bleeds). Same biology also means ε4 carriers run a higher baseline risk of spontaneous lobar hemorrhage even without these drugs, which is part of why blood thinners compound the danger.

Because of that gradient, regulators and guidelines have converged on a few points, summarized in the [NCBI genetics summary on lecanemab and APOE](https://www.ncbi.nlm.nih.gov/books/NBK605938/):

-   APOE genotyping is recommended before starting, so the ARIA risk is known and discussed up front.
-   Some regulators have treated homozygotes differently from other patients, reflecting their higher risk, so eligibility rules can depend on your genotype and where you live.
-   Treatment requires scheduled MRI monitoring, especially during the first year when ARIA is most likely.
-   Blood thinners and certain other factors further raise bleeding-related risk and feed into the decision.

## Where the approvals stand

The regulatory map is genotype-aware and splits by region, which directly affects who can get these drugs.

-   United States: the FDA granted lecanemab (Leqembi) full approval in July 2023 and donanemab (Kisunla) in July 2024, both with boxed warnings about ARIA and labeling that flags higher risk in ε4 homozygotes.
-   Europe: after an initial rejection, the European Commission approved lecanemab in April 2025, but only for ε4 non-carriers and heterozygotes, excluding ε4/ε4 homozygotes because of their ARIA risk. The EMA refused donanemab. So in the EU, a homozygous carrier currently cannot access these drugs at all, a stark example of genotype shaping eligibility.

For the adjacent pipeline, see [APOE-targeted therapies on the horizon](/topics/apoe-targeted-therapies-horizon). This area moves fast, so confirm the current rules where you live with a specialist.

## Early disease, genotype, and a specialist conversation

-   These drugs apply to early-stage, amyloid-confirmed disease under specialist care, not to APOE4 carriers in general and not as prevention.
-   Genotype affects both eligibility and safety, which is a concrete reason to learn your status with clinical guidance rather than in isolation. See [how to get tested](/topics/how-to-get-tested-for-apoe4).
-   For some ε4/ε4 individuals the risk-benefit balance is genuinely harder, and it becomes an individualized conversation with a specialist.
-   Approvals, eligibility criteria, and monitoring protocols keep changing, so verify current specifics with a qualified clinician.

## Common questions

**Can I take lecanemab if I’m APOE4 homozygous (ε4/ε4)?** It depends heavily on where you live. In the US, yes, with extra caution and a boxed warning, since homozygotes carry the highest ARIA risk. In the EU, the approval explicitly excludes ε4/ε4 homozygotes, so it is not available to them there. Either way it is a careful, individualized specialist decision.

**What exactly is ARIA, and would I feel it?** It is swelling or small bleeds visible on MRI. Often there are no symptoms, which is why monitoring scans matter. When symptoms do occur they can include headache, confusion, dizziness, or visual changes, and they should prompt immediate medical attention.

**Are these drugs a cure?** No. They clear amyloid and modestly slow decline in early disease. They do not restore lost function or stop the disease entirely.

If this decision is on the table for you or a family member, take it to a specialist who can weigh the modest benefit against the carrier-specific ARIA risk, especially for ε4/ε4.

## Sources & further reading

1.  [van Dyck et al. (2023), NEJM: Lecanemab in Early Alzheimer’s Disease (CLARITY AD)](https://pubmed.ncbi.nlm.nih.gov/36449413/)
2.  [Sims et al. (2023), JAMA: Donanemab in Early Symptomatic Alzheimer Disease (TRAILBLAZER-ALZ 2)](https://pubmed.ncbi.nlm.nih.gov/37459141/)
3.  [Alzheimer’s Association: Lecanemab (Leqembi)](https://www.alz.org/alzheimers-dementia/treatments/lecanemab-leqembi)
4.  [Alzheimer’s Association: Donanemab](https://www.alz.org/alzheimers-dementia/treatments/donanemab)
5.  [Medical Genetics Summaries (NCBI Bookshelf): Lecanemab Therapy and APOE Genotype](https://www.ncbi.nlm.nih.gov/books/NBK605938/)
6.  [Sun et al. (2025), J Prev Alzheimers Dis: Modified donanemab titration reduces ARIA-E (TRAILBLAZER-ALZ 6, 18-month results)](https://pmc.ncbi.nlm.nih.gov/articles/PMC12413699/)
7.  [Review (2025), PMC: APOE4, cerebral amyloid angiopathy, and ARIA risk](https://pmc.ncbi.nlm.nih.gov/articles/PMC11703089/)

## Related deep dives

-   [Valiltramiprosate (ALZ-801): The Oral Anti-Amyloid Drug Built for APOE4 Carriers What the APOLLOE4 Phase 3 trial of oral valiltramiprosate (ALZ-801) in APOE4/4 homozygotes actually showed: the miss, the MCI signal, no excess ARIA, and the p-tau217 data.](/topics/valiltramiprosate-alz-801-apoe4)
-   [HDL, ApoA-I, and Clearing Amyloid From the APOE4 Brain's Blood Vessels An ApoA-I Milano variant cut vascular amyloid in Alzheimer's mice. What HDL apolipoproteins mean for cerebral amyloid angiopathy in APOE4 carriers.](/topics/hdl-apoa1-cerebral-amyloid-clearance-apoe4)
-   [Blood-based biomarkers for Alzheimer’s: the coming shift For years, confirming Alzheimer’s biology meant a spinal tap or a PET scan. Blood tests are starting to change that. What they measure, where they stand, and the real caveats.](/topics/blood-based-biomarkers-alzheimers)