A new study from Duke Health suggests the Alzheimer’s drug lecanemab is performing in real-world settings much like it did in clinical trials. As the FDA-approved therapy moves into everyday care, researchers tracked more than 230 patients to see how safe the treatment actually is outside of tightly controlled environments.
The results, published in Neurology Open Access, offer some reassurance to families weighing the risks of side effects. The study focused on treatment persistence and disease progression, specifically looking for Amyloid Related Imaging Abnormalities, known as ARIA.
Safety Rates in the General Population
Researchers found that 79% of patients who started treatment were able to continue it for at least a year. The remaining patients who stopped mostly did so because of adverse events, though the majority of side effects resolved on their own without extra medication or hospital stays.
“There’s been a lot of fear about this medication, especially around brain swelling and bleeding,” says Andrew Liu, an associate professor at Duke University School of Medicine. “What we found is that, in the real world, with careful selection and close monitoring, approximately 80% of patients are able to remain on lecanemab for at least one year.”
Related: Common Plastic Linked to Higher Risk of Fatty Liver Disease
Data shows 24.3% of patients developed ARIA. Those with a specific high-risk Alzheimer’s gene, called APOE ε4 homozygotes, were about four times more likely to develop the condition than others. Regarding sex, rates did not differ.
ARIA events appeared as late as 30 weeks after starting therapy, with peaks observed around 10 and 25 weeks. The study also noted that 31% of patients experienced serious adverse events, though some were unrelated to the drug itself. Common issues included infusion reactions, falls, or strokes, often tied to the patient’s other underlying illnesses.
While the data is generally positive, the medical community still faces a significant challenge in predicting exactly who will suffer these complications before treatment begins. Current blood tests and imaging scans are not yet reliable enough to screen out high-risk individuals with total certainty, which means doctors must rely on careful vigilance rather than prediction.
A Patient’s View on Treatment
Sally Osmer was 74 when she received an early-stage Alzheimer’s diagnosis. Having a family history of the disease, she didn’t hesitate to start lecanemab infusions at Duke within a month of getting the news.
“When I got the diagnosis, it was pretty devastating,” Osmer says. “The prospect of having a slow decline but being physically healthy is very frightening. We were fortunate to catch it early, and I was pleased to begin treatment.”
Related: Message may prompt environmental cleanup
She has now completed more than a year on the therapy without serious side effects. Osmer says the treatment helps her stay engaged with her grandchildren and manage her anxiety about the future. “Being on lecanemab definitely gives me a sense of hope and a promise for a better future than we’ve assumed about Alzheimer’s in the past,” she says.
Monitoring and Future Needs
The Duke team piloted an imaging tool supported by artificial intelligence to help detect subtle ARIA changes earlier on brain scans. They plan to expand its use soon as part of their safety strategy. The center has treated more than 400 patients with lecanemab, making it one of the most experienced sites in the country using anti-amyloid therapies.
Researchers emphasize that cerebral amyloid angiopathy, a silent buildup of amyloid in blood vessels, is present in over half of all Alzheimer’s patients and is a key mechanism behind ARIA. P. Murali Doraiswamy, a study coauthor and professor at Duke, says the findings highlight a need for better predictive tests.
“Our study highlights the urgent need to develop better tests to predict who will develop serious ARIA and to know when it is safe to resume treatment,” Doraiswamy says.
Related: Edwards Sees Boost in Heart Valve Sales
The study found that seven commonly available clinical, blood, and imaging tests could not yet reliably predict who would develop ARIA. One method involving an MRI measure of vascular changes and blood levels of the biomarker pTau/AB42 showed the most promise for identifying low-risk subjects.
Liu curates a biorepository of blood and fluid samples from treated patients to help develop those biomarkers. He notes that ARIA status was not associated with changes in cognitive performance at 12 months.
The drug does not cure the disease or reverse symptoms, but researchers believe slowing progression allows patients to maintain independence. “This is about buying time in a disease that takes it away,” Liu says. “And our real-world experience shows that with careful risk-benefit assessments, time can be gained by many patients.”
Liu’s role was funded by the Ann B. Bussel Award and the Duke-UNC NIH award. Several study authors have relationships with pharmaceutical companies, including the manufacturer of lecanemab.