What ARIA is, and why it gates how anti-amyloid antibodies can be used

What it is

ARIA stands for amyloid-related imaging abnormalities. It is a class of brain-imaging findings that occur in some patients receiving anti-amyloid antibody treatment for early Alzheimer's disease. ARIA was identified in the clinical-trial era of the anti-amyloid antibody class and is now a defining feature of how the class is monitored and managed in routine care.

There are two recognised forms. ARIA-E refers to vasogenic edema or sulcal effusion - fluid accumulation in the brain that shows up as bright signal on FLAIR MRI sequences. ARIA-H refers to microhaemorrhages or superficial siderosis - small areas of bleeding or iron deposition that show up on susceptibility-weighted MRI sequences. The two can occur together or independently.

How it works

The leading scientific explanation is that anti-amyloid antibodies, by attaching to amyloid deposits in the walls of blood vessels (cerebral amyloid angiopathy) and clearing them, transiently increase vessel-wall permeability and fragility. The fluid that accumulates is the ARIA-E pattern; the small bleeds are the ARIA-H pattern. Both reflect the temporary stress that amyloid clearance puts on vessel walls.

ARIA can be entirely asymptomatic - the person feels nothing and the finding is detected only on routine surveillance MRI - or it can be symptomatic, ranging from mild headache or confusion to more serious neurological events that require treatment discontinuation and clinical management.

Who is more affected

Three factors consistently associate with higher ARIA risk:

The APOE genotype is the most important. People with two copies of the APOE4 allele (homozygotes, ~2% of the general population, ~15% of mild AD patients) have substantially higher rates of ARIA-E than people with one copy or none. Prescribing decisions in this subgroup require explicit benefit-risk discussion and genotyping is standard before treatment initiation.

The dose and intensity of treatment matter. Donanemab's titration protocol concentrates ARIA risk in the early infusions; lecanemab's continuous protocol distributes risk over time. The pooled rates differ between the two products partly because of these protocol differences.

The presence of cerebral amyloid angiopathy at baseline raises risk, which is why baseline MRI is required before initiation and patients with significant baseline microhaemorrhages are excluded from treatment.

Why it matters clinically

ARIA is the reason both products carry an MRI surveillance protocol on their FDA labels. Lecanemab's label calls for MRI before infusions 5, 7 and 14, with additional scans if symptoms develop. Donanemab's label calls for MRI before infusions 2, 3, 4 and 7. Beyond the labels, real-world neurology practice has settled into more frequent surveillance for higher-risk patients (APOE4 homozygotes especially) and looser cadence for lower-risk patients.

When ARIA is detected, the clinical decision depends on severity: mild asymptomatic ARIA may allow treatment continuation with closer monitoring; moderate ARIA usually requires temporary suspension; severe ARIA, particularly with neurological symptoms or large microhaemorrhages, typically ends the treatment course.

The MRI surveillance burden is a real operational and patient-experience cost. Each scan is a patient appointment, each scan costs the health system, and each scan competes for scanner time against other indications. The parallel commercial Signal piece on infrastructure rate-limiting covers the systems-level consequences.

What is changing

A current discussion in the field, with FDA engagement, is whether the surveillance MRI cadence can be relaxed for the lowest-risk patients (non-APOE4, no baseline microhaemorrhages, no early-treatment ARIA). Real-world data through 2025-26 shows that this subgroup has very low ARIA rates and that the marginal yield from frequent surveillance MRI is small. Less surveillance, if endorsed, would meaningfully ease the operational burden of the class.

The other shift to watch is the next generation of anti-amyloid agents (subcutaneous lecanemab, trontinemab, remternetug, and others in earlier development) that may have different ARIA profiles. If a next-generation agent shows materially lower ARIA rates while maintaining clinical efficacy, the class's operational footprint becomes much smaller and the addressable population grows accordingly.