ARIA stands for amyloid-related imaging abnormalities: brain-scan changes that can appear when someone takes an anti-amyloid Alzheimer's antibody. They show up on MRI as brain swelling (ARIA-E) or tiny bleeds (ARIA-H), and both Leqembi and Kisunla carry an FDA boxed warning about them, according to the Leqembi prescribing information. Most ARIA cases cause no symptoms and resolve, but a minority become serious and, rarely, fatal. This page explains what ARIA is, how the two approved drugs compare, who faces the highest risk, and what monitoring can catch problems early.
Medical information disclaimer: This article is for general educational purposes only and does not provide medical advice, diagnosis, or treatment. Always consult a physician or other qualified health professional about symptoms, medications, tests, or treatment decisions.
Table of Contents
- What ARIA actually is on a brain scan
- Leqembi vs. Kisunla — the ARIA numbers
- Who faces the highest risk?
- How monitoring and dosing lower the danger
- What these drugs can and cannot do
- Frequently Asked Questions
What ARIA actually is on a brain scan
ARIA is not one thing but two related findings. ARIA-E means edema — fluid swelling in brain tissue or effusions in the grooves of the brain. ARIA-H means hemorrhage — microbleeds or a surface iron staining called superficial siderosis. These changes come from anti-amyloid antibodies, drugs designed to clear the sticky amyloid plaque linked to Alzheimer's.
As the treatment strips plaque from blood-vessel walls, those vessels can leak or bleed, which is what the MRI picks up. The FDA labels list ARIA as the signature risk of this drug class. Many people never feel it. When symptoms do occur, they can include headache, confusion, dizziness, vision changes, nausea, or, rarely, seizures. Any of these during treatment is a reason to call the prescribing clinician promptly.
Leqembi vs. Kisunla — the ARIA numbers
Both drugs are IV infusions for early Alzheimer's, but their trial ARIA rates differ. leqembi (lecanemab-irmb) won traditional FDA approval on July 6, 2023, for mild cognitive impairment or mild dementia due to Alzheimer's, per the FDA approval announcement. Kisunla (donanemab-azbt) was approved July 2, 2024, for early symptomatic Alzheimer's.
In the CLARITY AD trial, 12.6% of people on Leqembi developed ARIA-E versus 1.7% on placebo, and most cases caused no symptoms, according to the Leqembi label. Kisunla ran higher: ARIA-E occurred in about 24% of treated patients in TRAILBLAZER-ALZ 2, roughly 6% with symptoms, and serious or fatal ARIA events did occur, per the Kisunla label. A quick comparison:.
- Leqembi: ARIA-E in 12.6% of treated patients; most asymptomatic
- Kisunla: ARIA-E in ~24%; about 6% symptomatic
- Kisunla difference: dosing can stop once amyloid plaque is cleared
Who faces the highest risk?
Your genes matter here. Both labels require ApoE ε4 genotyping before treatment because people who carry two copies of the ε4 gene — called homozygotes — have higher rates of symptomatic, serious, and severe ARIA, according to the FDA prescribing information.
Blood thinners raise the stakes too. ARIA can rarely turn serious or fatal, and the risk climbs for patients on anticoagulants or those who are ε4 homozygotes, as noted in FDA and NIH resources. Before starting either drug, patients and families can reasonably ask their clinician:.
- What is my ApoE ε4 status, and how does it change my risk?
- Am I on any blood thinner that increases bleeding risk?
- What symptoms should prompt an urgent call or MRI?
How monitoring and dosing lower the danger
MRI scheduling is the main safety net, and it has grown stricter for Leqembi. In late 2024 the FDA added an earlier MRI between the second and third infusions, plus scans before the fifth, seventh, and fourteenth infusions, to catch ARIA sooner, per the FDA drug safety communication. Kisunla took a dosing route.
On July 9, 2025, the FDA approved a modified titration regimen that lowered ARIA-E by 41% at 24 weeks and 35% at 52 weeks versus the original schedule, without losing plaque or P-tau217 efficacy, according to Lilly's announcement. Slower ramp-up appears to give blood vessels time to adjust. Neither approach removes the risk entirely. They shift the odds by finding ARIA before it becomes dangerous and by easing the body into treatment.
What these drugs can and cannot do
Set expectations before weighing the risk. These antibodies modestly slow cognitive decline; they do not reverse damage or cure Alzheimer's, as summarized in NIH and FDA resources.
That trade-off is the heart of the decision. A patient accepts real ARIA risk and a schedule of infusions and MRIs in exchange for a slower slide, not a recovery. For ε4 homozygotes or people on anticoagulants, the balance tilts differently than it does for a lower-risk patient, which is exactly why genotyping and a candid conversation with the prescriber come first.
Frequently Asked Questions
Does ARIA always cause symptoms?
No. Most ARIA cases found on MRI cause no symptoms and resolve, but a minority bring headache, confusion, or vision changes, and rare cases are serious or fatal.
Why do I need genetic testing before treatment?
Both labels require ApoE ε4 genotyping because people with two ε4 copies have higher rates of symptomatic, serious, and severe ARIA.
Can I take these drugs if I'm on a blood thinner?
Anticoagulants raise the risk of serious ARIA, so this needs a direct discussion with your prescriber before starting.
