Skip to main content
Premium Trial:

Request an Annual Quote

At AAN24, Researchers Seek to ID Risk Factors for ARIA After Anti-Amyloid Alzheimer's Treatments

3D illustration showing amyloid plaques in Alzheimer's disease

NEW YORK – Researchers have been exploring risk factors for developing a typically asymptomatic, but sometimes fatal, adverse event that's been linked with a growing class of anti-amyloid Alzheimer's medications: amyloid-related imaging abnormalities (ARIA). Two groups presented new data on possible risk factors at the American Academy of Neurology's 2024 annual meeting in Denver on Sunday.

There are two monoclonal antibody Alzheimer's treatments currently on the market targeting beta-amyloid, Biogen and Eisai's Leqembi (lecanemab-irmb), which gained full approval from the US Food and Drug Administration last summer, and Aduhelm (aducanumab-avwa), which is being discontinued later this year after a controversial approval in 2021 and slow sales. A third, Eli Lilly's donanemab, is under FDA review.

ARIA, which can present as temporary swelling or bleeding in the brain, has been one of the most common side effects seen in clinical trials of anti-amyloid drugs. While ARIA is usually asymptomatic, it can be associated with headache, dizziness, and nausea, and, infrequently, might involve serious brain edema and severe neurological symptoms. Intracerebral hemorrhages have also been observed in patients treated with anti-amyloid drugs, which can be fatal, according to the FDA.

The FDA required a boxed warning for Leqembi on ARIA, in which it recommends genetic testing for patients before beginning treatment, since patients with two copies of the APOE ε4 allele have shown a higher incidence of the condition.

APOE ε4 carrier status and how many microhemorrhages a patient has had before starting treatment are known variables associated with risk of developing ARIA, said Marwan Sabbagh, a professor in the neurology department at the Barrow Neurological Institute in Phoenix, in an interview.

"The question then becomes, is there anything else … that we should be exploring?" said Sabbagh, who presented a poster on ARIA risks during a poster session Sunday. "There are still a lot of unanswered questions."

In a subgroup analysis, Sabbagh and colleagues analyzed data from Eisai's randomized-controlled Phase III trial, CLARITY AD, which was the basis of the FDA's decision to convert Leqembi's accelerated approval to a traditional approval. They compared ARIA rates based on patients' baseline beta-amyloid levels, whether patients had mild cognitive impairment or mild Alzheimer's, and APOE ε4 status.

Overall, in the CLARITY AD trial, ARIA due to hemosiderin deposition (ARIA-H) occurred in 16.9 percent of treated patients and 8.9 percent of patients who received placebo. ARIA due to edema (ARIA-E) occurred in 12.6 percent and 1.7 percent of patients who received Leqembi and placebo, respectively.

The majority of ARIA cases were asymptomatic. Symptomatic ARIA-E was observed in just 2.8 percent of treated patients and none of the patients in the placebo group, and symptomatic ARIA-H was observed in 0.7 percent and 0.2 percent of Leqembi-treated and placebo patients, respectively, according to Phase III results previously published in the New England Journal of Medicine. There were no deaths that investigators determined were linked to Leqembi and none occurred with ARIA.

Initially, Sabbagh said he had expected that patients' baseline beta-amyloid levels would predict their probability of developing ARIA, with patients with more beta-amyloid likely to have more imaging abnormalities related to the protein. However, that wasn't the case. While patients with higher baseline beta-amyloid numerically had higher rates of ARIA, it wasn't statistically significant.

The subanalysis confirmed that APOE ε4 homozygous patients had the highest rates of ARIA when compared to APOE ε4 heterozygotes and non-carriers. However, neither baseline beta-amyloid levels nor disease stage had a significant effect on ARIA rates. In fact, frequency of ARIA-E and ARIA-H were similar within each APOE ε4 group, regardless of baseline beta-amyloid levels.

"The take-home message is that ARIA risk is independent of your amyloid load," Sabbagh said.

Sabbagh said it isn't clear why that is, and that additional research is needed. He's also curious whether baseline tau, another hallmark of Alzheimer's, could be associated with ARIA risk.

Eisai, which funded the study, in an emailed statement said that there are additional analyses ongoing that are investigating ARIA risk and that the company will continue to share updates.

"Prior to initiation of treatment, physicians should discuss with patients the risk of ARIA across genotypes and the implications of genetic testing results," a company spokesperson wrote. "Safety is a very important consideration, and the incidence and timing of ARIA vary among treatments."

In a separate analysis presented at the poster session on Sunday, researchers identified APOE genotype and certain factors identified on MRIs as variables associated with rates of ARIA-E in clinical trials of donanemab, Lilly's anti-amyloid monoclonal antibody candidate for Alzheimer's. Baseline beta-amyloid levels were also associated with ARIA-E, though not as strongly.

Lilly last year submitted an application to the FDA seeking traditional approval for donanemab, and last month said the FDA had indicated it would convene a meeting of the agency's Peripheral and Central Nervous System Drugs Advisory Committee to weigh in on the drug's safety and efficacy. Lilly did not immediately respond to a request for comment on the poster findings.

ARIA was one of the most common treatment-emergent adverse events within the randomized-controlled Phase III TRAILBLAZER-ALZ 2 trial of the drug, results of which were previously published in the Journal of the American Medical Association. ARIA was observed in 36.8 percent of treated patients and 14.9 percent of those receiving placebo and was more frequent in patients who carried two copies of the APOE ε4 allele. ARIA-E, specifically, was observed in 24 percent and 2.1 percent of treated and placebo patients, respectively.

Roughly three-quarters of ARIA cases were asymptomatic. However, in the donanemab group, three patients with serious ARIA died. Of the patients who died, two carried one copy of the APOE ε4 allele and the other was a non-carrier.

The analysis presented Sunday, a post hoc exploratory analysis of multiple clinical trials of donanemab, was sponsored by Lilly.

Researchers for the post hoc analysis used machine learning to analyze 42 variables, including patient characteristics, comorbidities, and concomitant medications, and whether they were related to ARIA-E risk. They identified APOE ε4 and, as measured by MRIs, baseline number of microhemorrhages and presence of cortical superficial siderosis as factors that increase risk of ARIA-E. Other factors like mean arterial pressure and beta-amyloid levels also had smaller contributions to ARIA-E risk, while ARIA-E rates decreased with use of antihypertensive medications.

"This analysis is hypothesis-generating for future validation work across the class of [amyloid-targeting therapies] and may support treatment decisions," researchers wrote.

The novel findings, like the influence of blood pressure and antihypertensive use on ARIA risk, need to be replicated in other studies, said Steven Greenberg, coauthor of the poster and professor of neurology at Harvard Medical School, in an interview. Additional research is also needed to better understand the impact of ARIA — and to what extent it's harmful even when asymptomatic — and to see if there are ways to reduce ARIA risk. For example, if validated, blood pressure could be a modifiable risk factor.

He stressed that ARIA risk is just one piece to weigh when deciding whether an anti-amyloid medication is appropriate for a patient, alongside the benefits of potentially slowing disease progression.

"That's going to be an individualized decision that will take in a lot of factors," Greenberg said. "We want to continue to get better at predicting risk of ARIA to help make that decision as informed as possible."