Leqembi in the Clinic: So Far, ARIA Echoes Trials

Series - AD/PD™ 2025: Advances in Science & Therapy:

As lecanemab gains regulatory approval in more countries (see previous story), clinicians are eager to learn from the first centers to have incorporated this therapeutic antibody into their clinical practice. At the International Conference on Alzheimer’s and Parkinson’s Diseases, held April 1-5 in Vienna, speakers from the U.S. and Israel described their experience treating the first several hundred patients in their centers to an engaged audience that peppered them with questions.

  • In the U.S., lecanemab treatment happens mostly in cities.
  • In Israel, APOE4 homozygotes are not treated, and ARIA-E rates are low so far.
  • As Europe prepares to offer treatment, its radiologists want more training on ARIA.

The scientists reported that so far, adverse events in clinical practice resemble those in the trials. In some settings, rates of ARIA-E are lower than expected, perhaps because many medical centers exclude APOE4 homozygotes. In support of withholding lecanemab from this group of patients, Stefan Teipel at the German Center of Neurodegenerative Diseases, Rostock, presented a Bayesian reanalysis of Clarity Phase 3 data by which homozygotes derived no benefit from lecanemab.

Meanwhile, lecanemab was approved in Europe on April 15, which will give many more people worldwide access to the drug. In Vienna, Tiago Gil Oliveira at the University of Minho in Braga, Portugal, reported that European clinicians feel uncomfortable managing ARIA and want training. Efforts to offer that are underway.

Tamara Shiner at the Tel Aviv Sourasky Medical Center said the advent of amyloid immunotherapy requires physicians to change how they think about Alzheimer’s treatment. Instead of prescribing symptomatic pills that require little urgency or follow-up, they now need to diagnose AD patients quickly, learn about infusions, and monitor their patients closely. Still, tertiary clinics are making the change. “It is feasible to give anti-amyloid antibodies, although it requires a lot of work and organization,” Shiner said in Vienna.

U.S. Use Still Limited
To get an overview of lecanemab clinical use in the U.S., Michael Rosenbloom at the University of Washington, Seattle, turned to the Purple Lab CLEAR Claims database, which gathers information on 330 million U.S. patients. In Vienna, Rosenbloom said that he searched the database for anyone who had received at least one dose of lecanemab and had at least six months of run-in data beforehand. This brought up 4,903 people, slightly more than a third of the estimated 13,500 who have taken the drug as of January 2025.

This group averaged 76 years old, five years older than the Clarity trial population, and 55 percent were women. Overall, 77 percent were white, 5 percent from underrepresented groups, and 17 percent of unknown race. About two-thirds had mild AD dementia, the rest mild cognitive impairment, suggesting patients were at a more advanced disease stage on average than in the trial, where those percentages were flipped.

Notably, 98 percent of patients came from urban areas. Rosenbloom noted that many rural areas remain “treatment deserts,” with few neurologists and little access to innovative therapies. In this dataset, 82 percent of prescribers were neurologists; 5 percent were in family practice.

Nearly all the patients started treatment after July 2023, when lecanemab received traditional approval from the Food and Drug Administration. It took 79 days on average before patients received their first dose, and the average time between doses was two weeks, as expected. About 80 percent stayed on lecanemab for at least six months, which is higher than the 50 percent adherence to acetylcholinesterase inhibitors, Rosenbloom said.

The data implied that physicians were following guidelines for managing ARIA, with the first MRI occurring on average 54 days after starting treatment, and about a fifth of patients having a treatment gap of 90 days or more, consistent with stopping infusions due to ARIA. However, the dataset offered no direct information on ARIA.

“The FDA label can be followed in the real world,” Rosenbloom concluded. Previous overviews of U.S. lecanemab treatment have reached the same conclusion (Nov 2024 conference news).

One U.S. Clinic’s Experience
In Vienna, Andy Liu at Duke University in Durham, North Carolina, described how treatment is going at his clinic. Duke began offering lecanemab two years ago, and the clinic has treated 223 people to date. With an average age of 74, they were three years older than the Clarity trial population. Nearly all were white and college-educated. Two-thirds were diagnosed with mild cognitive impairment; the rest with mild AD. Their average MMSE score was 26. A majority carried an APOE4 allele, with about half the population being E4 heterozygotes and another 10 percent homozygotes.

For these initial patients, it took a year before they received their first infusion, reflecting the challenges of setting up the infrastructure to deal with administering amyloid immunotherapy, Liu said. By now, most of these initial patients have been on the drug for a year; many have completed 18 months and have transitioned to maintenance dosing. This option, at one infusion per month, was recently approved by the FDA (Jan 2025 news).

Does lecanemab help patients? Liu said that most people did not notice a big difference after they started treatment. A few called it life-changing. “They feel less foggy, and are able to process information faster and concentrate better,” Liu said in Vienna. Alzheimer’s disease is heterogenous, with patients progressing at different rates and many having co-pathologies that accelerate their decline. Some researchers believe that people who have pure AD pathology and are caught early in the disease will fare best on amyloid immunotherapy. Liu said that as biomarkers for other pathologies become available, scientists should model lecanemab’s efficacy with and without the presence of co-pathologies.

Regarding safety, infusion-related reactions were the most common side effect, occurring in 22 percent of patients, similar to the 26 percent incidence in Clarity. Typical symptoms were fever, chills, and headache, and resolved quickly. Concerningly, almost 20 percent of Duke immunotherapy patients reported falling, double the rate in the trial. “Dizziness and falls should be on our radar,” Liu said in Vienna, noting that his clinic tells nurses in the infusion center to watch for this. One clinician privately told Alzforum that such falls might be related to lecanemab and contribute to poor clinical outcomes. The FDA label lists dizziness and problems walking as two symptoms of ARIA. A small number of patients experienced other health problems such as stroke, sepsis, or pneumonia; it is unclear if these were related to the drug.

What about ARIA? Fifteen percent of Duke patients developed microhemorrhages, about the same as the 14 percent in Clarity. Other types of ARIA were less frequent than in the trial. About 6 percent of Duke patients had ARIA-E, and 1 percent superficial siderosis, compared to 13 and 6 percent, respectively, in the trial. Stratified by genotype, the 22 or so APOE4 homozygotes at Duke had low rates as well, with 5 percent developing ARIA-E and 9 percent ARIA-H. This is less than in the trial, where about one-third of homozygotes experienced each type.

Liu said he is unsure why that is. His clinic follows the lecanemab AUR, and these recommendations changed how he manages patients (Apr 2023 conference news). In addition, Duke clinicians screen out anyone with signs of cerebral amyloid angiopathy on MRI scans before treatment, he said. CAA has been linked to much worse outcomes from immunotherapy (Aug 2023 conference news). Scientists at Duke are currently doing predictive modeling to find out what patient characteristics correlate with ARIA.

Liu noted that Duke hospitalizes any patient who develops ARIA symptoms, or ARIA that is graded moderate or severe on MRI scans. For mild asymptomatic ARIA, physicians stop infusions and monitor the patient. Despite these precautions, two patients at Duke died, one from a macrohemorrhage, one from severe ARIA-E. Both were APOE4 heterozygotes. “A lot of patients with moderate or severe ARIA are heterozygotes,” Liu cautioned. The Duke clinic has only recently begun administering donanemab, and so has no data on that yet.

Avoiding Serious Outcomes
Stephen Salloway at Butler Hospital in Providence, Rhode Island, updated the AD/PD audience on known deaths on amyloid immunotherapy to date. He knows of nine on lecanemab, and six on donanemab. In addition, one death each has been attributed to aducanumab and trontinemab, for a total of 17.

This is five more than Alzforum had previously covered. Three deaths had been reported in the Clarity OLE, two in lecanemab clinical practice, and five in the donanemab Phase 3 trial or OLE (Jan 2023 news; Aug 2024 conference news; Jun 2024 news). The aducanumab and trontinemab deaths were covered as well (Nov 2021 conference news; Nov 2024 conference news).  

Overall, Salloway said, six of the deaths were due to macrohemorrhages, eight to severe inflammatory ARIA-E, and three more to ARIA-E that was mistaken for a stroke in the emergency room and erroneously treated with thrombolytics. All of those who got ARIA-E, and half of those with a macrohemorrhage, had an APOE4 allele. Salloway noted that eight cases had shown clear warning signs before the patient succumbed, suggesting those deaths could possibly have been prevented if doctors had known what to look for.

Some Alzheimer’s clinicians recommend a recent publication from the American Society of Neuroradiology’s Alzheimer, ARIA, and Dementia Study Group for its guidelines on detecting and managing ARIA. The paper describes three key MRI sequences to use, as well as three common patient scenarios and how to handle them (Cogswell et al., 2025).

Israel’s Experience: No Homozygotes, Not Much ARIA-E
In this small country of 9.7 million people, clinicians have been administering lecanemab for a year and a half. In Vienna, Shiner said her center has screened 156 people. Of those, 29 people were disqualified for medical reasons. These included 14 who were APOE4 homozygotes, as clinicians there decided not to treat this high-risk group. Another 30 candidates changed their minds about wanting the drug, 12 because they lacked insurance coverage. Eleven more are considering or plan to start, but haven’t done so yet. That leaves 86 people who had started lecanemab infusions as of January.

Their average age was 72, almost two-thirds were women. Their MMSE average was 24; about half carried APOE4. Shiner noted that MRI scans at her hospital revealed a surprisingly large amount of brain atrophy, with global brain volume in the ninth percentile for these patient’s age group. “Possibly MCI is already too late to start treatment, if this is the level of neurodegeneration,” she said.

So far, 13 people have developed ARIA-H, 12 of them without symptoms. Most had microhemorrhages and continued treatment, but clinicians stopped lecanemab in one person who developed superficial siderosis, a sign of a brain bleed. Three people have had ARIA-E, two of them mild and asymptomatic, and one moderate and symptomatic. Shiner suggested this low rate may be due to excluding E4 homozygotes. To date, 17 people have stopped treatment, including four due to ARIA and three for financial reasons.

Overall, the clinical experience with lecanemab is consistent with trial data, Shiner concluded. Her center has just begun giving donanemab.

Talya Nathan, also at the Tel Aviv Sourasky Medical Center, noted that the clinic has had to innovate in order to manage lecanemab prescriptions. Because the clinic only has four full-time neurologists on staff, and one half-time, they developed a system for community neurologists and geriatricians to refer patients. After referral, the medical center does an in-house workup that includes APOE genotyping, MRI scans, and confirmation of amyloid positivity by cerebrospinal fluid or PET scan. Based on each case’s urgency, patients are triaged into either a fast track that will allow them to start treatment within a few weeks, or a slower track. An interdisciplinary team meets weekly to review cases and decide on treatments.

In answer to an audience question, Nathan said that some patients decline immunotherapy after being informed of the risks, but most choose to continue. The greatest barrier to treatment is cost, she said. Lecanemab treatment is not currently covered by Israel’s national health insurance, and only half the country’s people have private insurance. As in the U.S., current lecanemab patients tend to be well-off city dwellers.

On the Horizon—European Use
What about other countries? Patients in Japan have been taking lecanemab for about a year. There were no updates in Vienna, but Takeshi Iwatsubo at the University of Tokyo told Alzforum that about 6,800 people in his country had received lecanemab as of January. By now, that number might be 8,000, he said. Meanwhile, donanemab has been available for about five months, and perhaps 1,000 people have received infusions of that antibody.

In the Clarity trial, about 6 percent of participants living in Asia developed ARIA-E, half the rate of other areas (Chen et al., 2025). Iwatsubo said that this remains true in clinical practice, with the incidence rate around 5 to 6 percent as of Eisai’s last adverse event report. Japan does not restrict treatment of APOE4 homozygotes, and the reason for the lower rate there is unknown. The National Center of Neurology and Psychiatry in Japan has set up a registry to collect amyloid immunotherapy data and parse what factors lead to good or bad outcomes.

With the European Commission’s approval of lecanemab on April 15, the drug is now approaching routine care in many more countries. In Vienna, Oliveira reported the results of a survey of ARIA knowledge among European healthcare professionals conducted by the European Society for Neuroradiology and the European Academy of Neurology. Roughly half of the 422 respondents were radiologists, the rest neurologists. They came from across Europe, from both academic and clinical settings, Oliveira noted.

The result? Seventy percent of respondents were familiar with ARIA, 60 percent with its risk factors. About 70 percent were confident they would be able to identify ARIA-H, 60 percent ARIA-E. However, only 9 percent said they were very confident in their ability to grade the severity of ARIA. In addition, only 39 percent were confident they could manage ARIA, and just 42 percent said their institutions could. Perhaps unsurprisingly, 93 percent of respondents wanted more training.

To provide it, the ESN and EAN have hosted online webinars in 2024 and 2025, and have held training sessions at scientific meetings. They are developing an educational website.

In Europe, APOE4 homozygotes will be ineligible for lecanemab, and at AD/PD in Vienna, Teipel showed data supporting this decision. He reanalyzed data from the Phase 3 trials of lecanemab and donanemab using Bayesian statistical methods, which allow scientists to calculate the probability of a hypothesis being true or false. For lecanemab, the analysis found five times more evidence for the absence of a clinical benefit in APOE4 homozygotes than for its presence. For the donanemab Phase 3 trial data, there was three times more evidence for the absence rather than presence of a benefit (Teipel et al., 2025). “Excluding APOE4 homozygotes from treatment is reasonable,” Teipel concluded, validating the compromise the U.K. and EU agencies struck. However, he cautioned that there could be confounding factors in the Phase 3 data, such as lower drug exposure for APOE4 homozygotes as a result of more frequent dosing pauses due to ARIA, and noted that more research is needed.—Madolyn Bowman Rogers

Comments

No Available Comments

Make a Comment

To make a comment you must login or register.

References

Therapeutics Citations

  1. Leqembi
  2. Kisunla
  3. Aduhelm
  4. Trontinemab

News Citations

  1. Leqembi: Side Effects No Worse in Clinical Use Than They Were in Trial
  2. Leqembi Maintenance Dosing Approved in the U.S.
  3. Next Goals for Immunotherapy: Make It Safer, Less of a Hassle
  4. Is ARIA an Inflammatory Reaction to Vascular Amyloid?
  5. Should People on Blood Thinners Forego Leqembi?
  6. Two New Deaths on Leqembi Highlight Need to Better Manage ARIA
  7. Donanemab Approval Likely to Pose New Quandaries for Clinicians
  8. Aduhelm Lowers Tau; Registry to Track Real-World Performance
  9. Trontinemab Data Strengthen Hope for Brain Shuttles

Paper Citations

  1. Cogswell PM, Andrews TJ, Barakos JA, Barkhof F, Bash S, Benayoun MD, Chiang GC, Franceschi AM, Jack CR Jr, Pillai JJ, Poussaint TY, Raji CA, Ramanan VK, Tanabe J, Tanenbaum L, Whitlow CT, Yu FF, Zaharchuk G, Zeinah M, Benzinger TS, ASNR Alzheimer, ARIA, and Dementia Study Group. Alzheimer Disease Anti-Amyloid Immunotherapies: Imaging Recommendations and Practice Considerations for Monitoring of Amyloid-Related Imaging Abnormalities. AJNR Am J Neuroradiol. 2025 Jan 8;46(1):24-32. PubMed.
  2. Chen C, Katayama S, Lee JH, Lee JY, Nakagawa M, Torii K, Ogawa T, Dash A, Irizarry M, Dhadda S, Kanekiyo M, Hersch S, Iwatsubo T. Clarity AD: Asian regional analysis of a phase III trial of lecanemab in early Alzheimer's disease. J Prev Alzheimers Dis. 2025 May;12(5):100160. Epub 2025 Apr 5 PubMed.
  3. Teipel S, Tang Y, Khachaturian A. Clinical efficacy of anti-amyloid antibodies in apolipoprotein E ε4 homozygotes: A Bayesian reanalysis of lecanemab and donanemab phase 3 results. Alzheimers Dement (N Y). 2025;11(2):e70083. Epub 2025 Apr 9 PubMed.

Other Citations

  1. previous story

External Citations

  1. Purple Lab CLEAR Claims
  2. webinars
  3. educational website

Further Reading

Annotate

To make an annotation you must Login or Register.