After a favorable review from a Food and Drug Administration advisory committee, Eli Lilly’s donanemab appears poised for marketing approval in the U.S. (Jun 2024 news). Are clinics ready? While last year’s approval of lecanemab is helping pave the way, donanemab’s unique aspects raise fresh questions. The June 10 AdComs grappled with issues such as whether tau PET scans are needed to identify the patients most likely to benefit, how clinicians will know when a patient has had enough donanemab, and if and when to put them back on. The committee also debated safety issues for certain subgroups. Due to a lack of data, they offered no definitive guidance on these points, and suggested leaving final decisions up to doctors and patients.

  • AdComs recommends not requiring tau PET before prescribing donanemab.
  • Guidance for when to stop dosing, and maybe restart, is unclear.
  • Donanemab has a more convenient dosing schedule than lecanemab, and more ARIA.

“The advisory committee’s votes will have the effect of making donanemab’s label nearly interchangeable with lecanemab’s, except for a more liberal, lower MMSE threshold,” predicted Lon Schneider at the University of Southern California, Los Angeles (comment below).

This means that for physicians, integrating donanemab into their practice may be akin to setting sail while finishing building the ship. Still, most clinicians Alzforum contacted welcomed the additional treatment option. They believe lecanemab and donanemab will complement each other, whereby lecanemab has a slight edge on safety and donanemab on less burdensome dosing. Each has advantages that will appeal to specific patients, researchers said.

Donanemab’s biologics licensing application rested on results from its single Phase 3 trial, with supportive evidence from Phase 2 (Jul 2023 conference news; May 2023 news; Mar 2021 conference news). In people with low to moderate tangle loads, the antibody slowed cognitive decline on multiple measures by about a third, while rapidly clearing amyloid. Because this efficacy was similar to lecanemab’s, many industry observers had expected the drug to be approved without an AdComs. FDA said it convened the committee to get expert input on donanemab’s label (May 2024 news).

The committee included three permanent voting members: ALS researcher Merit Cudkowicz at Massachusetts General Hospital, Boston; neuropathologist Thomas Montine at Stanford University School of Medicine, California; and Parkinson’s researcher Tanya Simuni at Northwestern University, Chicago. The other eight were temporary voting members: geriatrician Cynthia Carlsson at the University of Wisconsin, Madison; neurologist Nilüfer Ertekin-Taner at the Mayo Clinic in Jacksonville, Florida; biostatistician Dean Follmann at the National Institute of Allergy and Infectious Diseases, Bethesda, Maryland; neurovascular researcher Costantino Iadecola of Weill Cornell Medical College, New York; movement disorder specialist Kathleen Poston at Stanford; and neurologist Daniel Press at Beth Israel Deaconess Medical Center, Boston, as well as consumer representative Sarah Dolan at the Critical Path Institute in Tucson, Arizona, and patient representative Colette Johnston. Montine chaired the group. Conflict-of-interest rules tend to exclude many clinician-researchers who are actively involved in AD therapy development, particularly of anti-amyloid antibodies.

To Scan or Not to Scan?
What did this group recommend? Their first question was whether tau PET scans should be required for clinical use of donanemab. Trial participants were screened by tau PET, with those between 1.10 and 1.46 SUVr reaping the greatest benefits. Those above 1.46 fared less well, and those below 1.10 were enrolled into an open-label safety addendum that did not track cognition. About 8 percent of amyloid-positive people with cognitive impairment have a tangle load below 1.10, Lilly’s John Sims noted.

In the FDA’s presentation to AdComs, efficacy reviewer Kevin Krudys showed an analysis that broke out the high-tau group by itself. He reported that donanemab slowed progression in these patients by about 20 percent on the CDR-SB, and only 6 percent on the iADRS. This compares with 36 and 35 percent, respectively, for those with low to medium tangle loads. He also noted that biomarker changes in people below 1.10 SUVr, known as the “no-tau” group, were consistent with those in other groups, hinting at similar effects on the disease. “It’s reasonable to generalize [efficacy] across the spectrum of tau,” he concluded. This means patients would be selected based only on amyloid positivity combined with cognitive impairment in the MCI to mild dementia range.

This concerned the committee. Montine noted that the efficacy at higher tangle loads implies that people with advanced disease may hardly benefit, complicating a clinician’s decision whether to treat and when to stop. On the other end of the scale, Follman said he was uncomfortable extrapolating efficacy findings to the no-tau group based on biomarker readings alone, and suggested waiting for data from Trailblazer-Alz3, an ongoing secondary prevention study of donanemab in preclinical AD (Jul 2021 news; Nov 2021 conference news).

Ultimately, however, the committee felt requiring tau PET for selecting patients would be impractical. “Access to tau PET would be impossible for many communities,” Carlsson said, and most agreed. Poston noted that tau PET scans are only available at tertiary care centers. They recommended the label not demand this.

Cudkowicz asked whether a blood test might be able to detect high tangle loads. Lilly’s Mark Mintun noted that plasma p-tau levels correlate with tau PET at about r=0.7, but said the marker does not accurately predict tangles among amyloid-positive people. Nonetheless, Eric Siemers at Acumen Pharmaceuticals thinks this approach could hold potential for the future. “Future studies might use a plasma tau measure such as p-tau217, since these measures may be actually more sensitive than tau PET,” he wrote (comment below). Several candidate fluid tests of brain tangle load are being developed, but none are ready yet for this purpose (Aug 2023 conference news; Mar 2024 news).

In comments to Alzforum, Gil Rabinovici at the University of California, San Francisco, concurred with the committee’s recommendation not to require tau PET, but also urged the community to push to make both amyloid and tau PET more available, affordable, and covered by payors. “PET scans are used extensively to guide clinical decision-making in oncology, and we should not accept a lower standard in AD, as the field enters a new era of precision medicine,” he wrote to Alzforum (comment below). David Knopman at the Mayo Clinic in Rochester, Minnesota, noted that, because administering donanemab already requires deep multidisciplinary expertise, tau PET is likely to be accessible at many academic centers that will pioneer its use during its ramp-up phase, hence may be more available than the committee thought (see comment).

When to Stop, and Whether to Resume
Perhaps the most puzzling issue for the committee was how clinicians are to know when to stop treatment. Because donanemab targets a form of Aβ found only in plaques, Lilly recommends discontinuing dosing once plaque has been cleared. In Phase 3, participants were switched to placebo if their plaque load fell below 11 centiloids, or below 25 in two consecutive scans. In clinical practice, obtaining multiple amyloid PET scans would be cumbersome and expensive. How are clinicians to know when to test?

Quizzed by the committee on this point, Lilly’s David Hyman called the one-year mark a reasonable timepoint. In the trials, two-thirds of participants were amyloid-negative by then. Teresa Buracchio of the FDA noted that amyloid positivity could be determined by a simple visual read of the scan.

The committee considered this guidance inadequate. “When would you test again? How many times do you retest? I would want to have answers to implement this in the clinic,” Poston said. “A main point that emerged [from the AdComs discussion] is the complexity of handling the discontinuation of donanemab … I found this the murkiest aspect of an otherwise reasonable endorsement of the agent,” Dennis Selkoe at Brigham and Women’s Hospital, Boston, wrote to Alzforum.

Committee members asked how well cognitive benefits are maintained after stopping. In the 18-month trial, participants who completed their treatment course around the one-year mark continued to diverge from those on placebo over the next six months. The committee said more follow-up will be needed to find out what happens over longer timeframes, and whether patients should go back on donanemab if their decline picks up pace again.

“There’s a huge cost savings for stopping, provided that stopping is supported by long-term post-marketing data,” Simuni said.

Ertekin-Taner suggested that physicians check in with their patients regularly after dosing stops to monitor for worsening status that could trigger a decision to re-initiate treatment. Dolan asked that Lilly provide guidance to patients on what symptoms to watch for. At the moment, no such guidance exists. Lilly has calculated that it would take five years for patients who have completed treatment to become amyloid-positive again, but has not suggested when to restart treatment (Jul 2023 conference news).

A recent paper hints that the benefits of immunotherapy depend on more than simply clearing existing plaque. In a study of an amyloidosis mouse model given aducanumab, researchers led by John Fryer at the Mayo Clinic in Scottsdale, Arizona, found that treatment prodded microglia into an amyloid-devouring mode. After treatment stopped, microglia maintained this beneficial phagocytic phenotype for 3.5 months, but lost it by seven months (Cadiz et al., 2024). Donanemab has also been shown to clear plaque by stimulating microglia to gobble it up.

How to Minimize Risk?
As with other antRi-amyloid antibodies, safety was a major concern. Donanemab produces about twice as much ARIA-E as does lecanemab, with edema in a quarter of patients. About a fourth of them, i.e., 6 percent of all patients, develop symptoms. While serious complications are rare, three deaths in the Phase 3 trial, and two in the open-label extension, were linked to ARIA. Lilly’s Melissa Veenhuizen said that four of these patients carried one copy of the APOE4 allele. The fifth did not, but had superficial siderosis, iron deposits that flag previous microhemorrhages, in his brain at baseline. The main risk factors for ARIA-E are APOE4, superficial siderosis, or more than one microhemorrhage at baseline. The latter two are both signs of possible cerebral amyloid angiopathy (CAA), an otherwise invisible vascular disorder associated with worse outcomes from amyloid immunotherapy (Aug 2023 conference news).

Could these deaths have been prevented? Veenhuizen noted that about a third of ARIA-related serious adverse events occurred at the second monthly infusion, and almost another third at the third infusion. Very few occurred after six months. At the beginning of the Phase 3 trial, MRIs were done before the first, fourth, and seventh infusions. During the trial, Lilly added an MRI before the second infusion, enabling clinicians to halt dosing earlier if ARIA had developed. This cut the number of serious ARIA cases by a quarter, and symptomatic cases by a third, Veenhuizen said. Lilly recommends that the FDA label specify an MRI be added before the third infusion.

“We think additional MRIs will minimize risk. But we can’t entirely eliminate it,” Hyman told the committee. Reisa Sperling at Brigham and Women’s Hospital, who co-leads AHEAD, a secondary prevention study of lecanemab, told the AdComs that ARIA appears tied to both antibodies’ mechanism of action, hence its risks can only be mitigated. The company plans to offer training for physicians, as well as information cards patients can carry with them to inform health-care professionals they encounter of their donanemab use. This could help physicians spot signs of ARIA and avoid prescribing contraindicated medications.

The committee focused on the risk for APOE4 homozygotes. This group is more susceptible to ARIA than others, with more than half of them getting it on donanemab, versus 20 percent on placebo. In addition, this subgroup appeared to gain less clinical benefit from donanemab in the Phase 3 trial, Press noted. Follman disagreed, noting the number of homozygotes was small, and the group was not powered to show a clinical effect. Nonetheless, all members agreed the higher risk of ARIA worsens the risk-benefit ratio for homozygotes. Ertekin-Taner noted the importance of communicating this to people considering donanemab. At the same time, the committee felt APOE genotyping should be recommended, not mandated, and that the final decision be left up to patients and physicians. Buracchio noted that as with lecanemab, donanemab’s label will strongly recommend APOE genotyping.

Other risks were less clear. Very rarely, Alzheimer’s patients develop intracerebral hemorrhages, which are bleeds greater than 1 cm in size. Whether ICHs are more common on donanemab than placebo is unknown, due to a low incidence in the Phase 3 study, the FDA noted. Overall, there were three on donanemab, versus two on placebo, out of 1,000 patients in each group.

One specific situation is particularly hazardous. One person died after developing stroke-like symptoms while on donanemab and getting a thrombolytic drug in the hospital. This precipitated multiple intracerebral hemorrhages throughout his brain. Importantly, the FDA noted, ARIA symptoms can mimic stroke, leading to inappropriate treatment and catastrophic outcomes. Therefore, emergency room staff must be informed to use caution before administering clot-busting drugs to patients on amyloid immunotherapy. This is already the protocol at many hospitals where lecanemab is in use.

Stephen Salloway at Butler Hospital in Providence, Rhode Island, noted that the accumulating evidence indicates ARIA-related deaths occur in one of three ways: an ICH in a person with CAA; inflammatory ARIA-E that resembles CAA-related inflammation; or ARIA-E that mimics stroke and is mistakenly treated with thrombolytics. “These risks can be substantially mitigated by optimal patient selection and careful management of ARIA,” he wrote to Alzforum (comment below).

What about chronic use of blood-thinning drugs, i.e., anticoagulants and antiplatelet agents like aspirin? These medications were permitted in the Phase 3 trial but, other than aspirin, their use was rare. More than 300 patients took aspirin, while 91 took a different antiplatelet agent, and 98 took an anticoagulant. The medications did not appear to be associated with increased risk of ARIA-E, ARIA-H, or ICHs, but because of the limited numbers, the data remain inconclusive, the FDA noted.

In addition to these issues, donanemab causes infusion-related reactions in about 8 percent of patients. Typical symptoms are flushing, chills, nausea, and sweating. Most IRRs occurred during infusion or within 30 minutes afterward, and cleared up the same day. In three cases, or 0.3 percent of participants, the person developed anaphylaxis, a life-threatening immune reaction. Hyman noted that unlike for some anti-amyloid antibodies, administering anti-inflammatory drugs before infusing donanemab does not help prevent IRRs. Instead, Lilly recommends monitoring patients for 30 minutes after each infusion.

Lilly has proposed three post-market safety studies to further examine hypersensitivity reactions and the effects of antithrombotic drugs. One would be a registry study in the U.S., one a cohort study in the U.S., and the third a cohort study in Europe.

The committee lamented a lack of data on the risks and benefits of donanemab for certain populations, including underrepresented minorities, people with autosomal-dominant AD, and those with Down’s syndrome. Only 8 percent of Phase 3 participants were non-white, a persistent problem in Alzheimer’s studies. None had ADAD or Down’s. All of these groups might seek clinical care with donanemab. However, the committee opted not to recommend restrictions to the label. Hyman noted that Lilly is planning to launch studies in both ADAD and Down’s in the near future.

Donanemab or Lecanemab?
Overall, the committee believed that donanemab would be a valuable addition to the clinic. Press liked that donanemab is easier on the patient in that it requires monthly instead of twice-monthly dosing, and half an hour to infuse instead of an hour. This and the time limit on the treatment protocol could free up infusion center capacity and allow more patients to get treated, he said. Dolan thought the short course of treatment might motivate patients to stick with it.

Erik Musiek at Washington University in St. Louis predicted that adding donanemab to clinical practice would be relatively straightforward. “Patient screening criteria and the infrastructure and expertise needed to administer donanemab is likely to be the same as that needed for lecanemab, and monitoring for adverse events should be similar,” he wrote to Alzforum (comment below).

If both antibodies are available, which one should patients take? That depends on the patient, AD researchers stressed. “It is possible that lecanemab may be a better choice for higher-risk patients, such as APOE4 homozygotes, though real data will be needed to conclude this,” Musiek noted. Meanwhile, donanemab may make treatment more feasible for some patients. “Donanemab may be a good option for patients who are willing to accept those risks but live far away from an infusion center and would prefer dosing every four weeks,” Irina Skylar-Scott at Stanford wrote to Alzforum.

In the afternoon, 10 patients or caregivers who participated in donanemab trials spoke to the AdComs, all in favor of approval. Several ascribed remarkable improvements to the drug. One said her husband has become more talkative, walks with a normal stride again, and showers and gets dressed by himself once more. Another said her husband has gone from “sundowning,” i.e., becoming confused in the early evening, every day to once every six weeks. “Our experience in the trial has been life-changing,” she told the committee.—Madolyn Bowman Rogers


  1. The Perfect Meeting

    Three years ago, an FDA adviser called the agency’s approval of Aduhelm “probably the worst drug approval decision in recent U.S. history.” By contrast, a current adviser called today’s donanemab session “a perfect meeting.”

    FDA got the cover it needed from a well-orchestrated, orderly, and somewhat performative meeting. There was never doubt about donanemab’s eventual marketing approval. The Trailblazer-Alz2 Phase 3 trial clearly met last year’s established "Leqembi criteria" for regular approval. Rob Califf, the FDA administrator, said as much at the last JPM conference.

    Today’s advisory committee was about what to put in the label, the prescribing information, how to use donanemab. There was little discussion about the clinical meaning of the outcomes, which were about the same as Leqembi and aducanumab and no less controversial. Rather than discuss mean differences in outcomes, FDA and Lilly reflected donanemab’s effect in smoother-sounding relative terms, such as 29 percent or 36 percent slowing, or five months “time savings” compared to placebo.

    There were three key outcomes that must have delighted both Lilly and FDA. The committee agreed that: 

    1. Tau PET is not needed to qualify a patient for donanemab treatment as it was for the clinical trials.
    2. Reduction in plaques, plasma ptau217, and plasma GFAP are good enough, standalone surrogate markers for clinical benefit. An efficacy trial for patients who are amyloid-PET-positive, but tau-PET-negative is not needed.
    3. Amyloid PET scans are not needed to monitor treatment at six and 12 months as they were in the Phase 3 trial, yet still could be used to determine when to stop.

    The advisory committee’s votes have the effect of making donanemab’s label nearly interchangeable with lecanemab’s except for a more liberal, lower MMSE ≥ 20 threshold instead of lecanemab’s MMSE ≥ 22.

    Safety, Black Box warnings, the 1 percent SAE or hospitalization rate for ARIA, potential for deaths, and brain volume loss got little attention. Only at the end did committee members broach a discussion of safety, treating APOE4 homozygotes, structural barriers to treatment, challenges of interpreting MRIs and amyloid PETs, limited infrastructure, cost, and inequity.

    Finally, one advisor suggested that use of donanemab should be several times more efficient than lecanemab. With once-a-month dosing instead of twice, half the infusion time, and most treatment courses lasting about 12 months for plaque density to revert to normal, donanemab might require only a fraction of the commitment and a quarter of the medical infrastructure than Leqembi requires.

  2. The FDA advisory committee meeting held on June 10 for donanemab was another important advance for patients, families, and the field. Using the iADRS as the primary outcome variable, and the CDR-SB as a key secondary, the committee unanimously voted 11- 0 that the data supported clinical efficacy for donanemab. For the Phase 3 TRAILBLAZER-ALZ2 study (Sims et al., 2023), statistical significance was reached for both the iADRS and CDR-SB. For the smaller Phase 2 TRAILBLAZER-ALZ trial, using MMRM, statistical significance was achieved using the iADRS but not the CDR-SB (Mintun et al., 2021). The second vote by the committee on whether the benefit risk ratio was favorable was also unanimous in support of donanemab. Thus, the traditional approval of donanemab is very likely, meaning that traditional approval for lecanemab and donanemab will increase choices for patients and their families, as well as prescribing physicians.

    Despite this progress, there were many caveats discussed by the committee, and labeling for donanemab is likely to be complex. For the Phase 2 TRAILBLAZER-ALZ trial, the inclusion/exclusion criteria included a requirement that “low/medium” tau was present, defined using flortaucipir tau PET. For the Phase 3 TRAILBLAZER-ALZ2 study, the primary outcome was in the “low/medium” tau patients, but patients with “high” tau were not excluded so the entire population could be evaluated as a secondary analysis. No patients with tau PET below the “low/medium” criteria were allowed in either trial, until this population was added to an open-label extension to the Phase 3 study. There was a general consensus among the committee members that obtaining tau PET in clinical practice is essentially not feasible, and exploratory analyses showed a clinical effect of donanemab across all three tau groups, so the committee generally felt that a tau PET requirement should not be included in the label. Interestingly there was a question from the committee on whether a plasma tau measure could be used as a substitute for tau PET, but the Lilly response was that plasma tau measures and tau PET did not correlate well. Future studies might use a plasma tau measure such as pTau217, since these measures may be actually more sensitive than tau PET.

    During the Phase 2 and Phase 3 studies, treatment could be changed from donanemab to placebo in a blinded manner if subsequent amyloid PET scans in the protocol showed that patients became “amyloid negative.” Amyloid PET scans were obtained at baseline, six months, 12 months, and 18 months (endpoint). How this would be performed in practice was a matter of some discussion at the advisory committee meeting. Based on plaque lowering in the Phase 3 trial, Lilly suggested that a follow-up amyloid PET scan could be obtained in practice at about one year. Some clinical cognitive data were presented for people who stopped treatment with donanemab, but very little data on clinical outcomes (not simply amyloid load based on PET) in these patients, for instance two years after stopping drug, is available. Regardless of the exact label language that is used, payors may require that treatment is stopped after some period of time, perhaps after an amyloid PET scan, despite the paucity of clinical data regarding this approach. Amyloid PET scans allow quantification of neuritic plaque load, but do not assess the more toxic soluble Aβ species including protofibrils and oligomers.

    Finally, there was a great deal of discussion about the safety of donanemab, in particular with regard to ApoE4 status. Genotyping for ApoE is likely to be recommended in the label, and the additional risk of ARIA-E, and in particular symptomatic ARIA-E, should be discussed with patients and families. Whether payors might limit reimbursement for patients who are ApoE4 homozygotes is unknown. Multiple surveillance MRIs are likely to be recommended in the label, especially early in treatment, which will add to patient burden and out of pocket expenses. It is important to note that not all monoclonal antibodies behave identically, and different monoclonal antibodies appear to have different risks of ARIA-E. For donanemab, 40.6 percent of ApoE4 homozygotes developed ARIA-E in the Phase 3 trial. For lecanemab, which targets protofibrils but also reduces plaque, the rate of ARIA-E for ApoE4 homozygotes in a Phase 3 trial was 32.6 percent (van Dyck et al., 2023). For sabirnetug, a monoclonal antibody developed to selectively target soluble globular Aβ oligomers, in a recent Phase 1 study, none of six ApoE4 homozygotes developed ARIA-E (manuscript in preparation). This possible lack of ARIA-E risk in ApoE4 homozygotes with sabirnetug is now being evaluated in a Phase 2 study.

    Despite these complexities, patients with early symptomatic AD and their families now have lecanemab available as a disease-modifying treatment given traditional approval by FDA, and in the immediate future donanemab will likely be added to the armamentarium. While these new treatments are certainly not cures for AD, they represent the beginning of a new era in the management of this complex disease. 


    . Donanemab in Early Symptomatic Alzheimer Disease: The TRAILBLAZER-ALZ 2 Randomized Clinical Trial. JAMA. 2023 Aug 8;330(6):512-527. PubMed.

    . Donanemab in Early Alzheimer's Disease. N Engl J Med. 2021 May 6;384(18):1691-1704. Epub 2021 Mar 13 PubMed.

    . Lecanemab in Early Alzheimer's Disease. N Engl J Med. 2023 Jan 5;388(1):9-21. Epub 2022 Nov 29 PubMed.

  3. I commend the FDA PCNS Advisory Committee on a very thoughtful discussion today. I completely agree with the committee’s favorable conclusions regarding donanemab’s efficacy and risk/benefit profile. I also agree that the innovative and forward-thinking trial design raises some challenges as we think about real-world implementation, particularly regarding the use of PET scans.

    I agree with many on the committee who felt that targeting amyloid plaque clearance as measured by PET to limit the duration of therapy represents a major advance that will be important to implement in clinical practice to the degree possible. I also think that the trial data demonstrate that tau PET staging provides important information about likelihood and magnitude of clinical response that can inform the risks vs. benefits discussion with patients. As the treating neurologist, there are certainly cases where knowing tau PET stage would inform my recommendations, in conjunction with other clinical data. I completely agree with the committee that it is not currently feasible to require tau PET for everyone, and that repeated amyloid PET scans will also pose a challenge in many care settings. At least for the moment, it does not seem that our current fluid biomarkers are capable of reliably determining who has cleared amyloid, or of staging tau spread at the single-patient level. That said, my hope is that, rather than accepting a “lowest common denominator” approach to care, we strive as a field to elevate care. This should include efforts to enhance access, affordability and coverage for amyloid and tau PET.

    PET scans are used extensively to guide clinical decision-making in oncology, and we should not accept a lower standard in AD, as the field enters a new era of precision medicine.

  4. The FDA Advisory Committee’s unanimous recommendation that donanemab receive full approval represents an important step forward in the treatment of Alzheimer’s disease. ARIA, the principal adverse event, is usually transient, asymptomatic, and manageable early in treatment, but ARIA can be serious and fatal and AD experts have a collective responsibility to ensure that amyloid-lowering monoclonal antibodies are given safely.

    The main risk factors for ARIA are ApoE4 gene copy number, the dose of the drug, and the severity of cerebral amyloid angiopathy. Though the rate of ARIA-E varies with each antibody, serious episodes of ARIA occur in approximately 1.5 percent of treated patients across antibody programs, and the primary goal is to limit the number of cases with long-term consequences. ARIA-related deaths can be divided into three categories: intracerebral hemorrhage in the setting of antibody treatment and CAA, severe inflammatory ARIA-E resembling CAARI, and ARIA presenting as a stroke mimic with fatal intracerebral hemorrhage following thrombolytic treatment. These risks can be substantially mitigated by optimal patient selection and careful management of ARIA.

    The following two examples based on fatal cases reviewed at the advisory committee illustrate important mitigation strategies. In case 1, a patient had an episode of severe symptomatic ARIA-E at low dose and treatment was suspended. Dosing was uptitrated after edema resolved despite the presence of 22 incident microhemorrhages. Had dosing been permanently discontinued based on guidelines in the Lecanemab Appropriate Use Recommendations, the fatal episode of ARIA-E would likely have been avoided. In case 2, a patient presented to the emergency room with focal signs suggestive of a stroke. The non-contrast head CT showed no hemorrhage and a CTA was negative for large vessel occlusion. Tenecteplase was ordered with subsequent fatal hemorrhage. Had an MRI been obtained earlier, this episode of severe ARIA-E could likely have been safely managed with high-dose corticosteroids and careful monitoring.

    Treating clinicians, radiologists and emergency teams need to be trained to detect and manage ARIA. Additional safety MRIs, patient safety cards, notifications in the medical record and updates to stroke guidelines can alert clinicians that patients are taking an amyloid-lowering antibody and advise caution on the use of thrombolytics. Automated imaging software can be utilized to assist radiologists and clinicians in detecting and monitoring ARIA. Systematic tracking of serious outcomes is required through treatment registries such as ALZ-NET.

    Disclosure: Dr. Salloway has been a site PI and safety monitor for trials of lecanemab, aducanumab, donanemab, and gantenerumab. He is an author on the Phase 2 and Phase 3 studies of donanemab and the Appropriate Use Recommendations for lecanemab and aducanumab. He has been a consultant to Eisai, Biogen, Lilly, and Roche. He owns no stocks and has no patents related to AD drug development.


    . Lecanemab: Appropriate Use Recommendations. J Prev Alzheimers Dis. 2023;10(3):362-377. PubMed.

  5. I am pleased to hear the positive news on donanemab and agree that it is likely to be approved soon. Donanemab has some potential advantages over lecanemab, but also may have some weaknesses, so it will be nice for clinicians to have both as options. Certainly the monthly dosing of donanemab is an advantage over the current two-week dose frequency for lecanemab. The defined timeframe of treatment is also a potential advantage and could save patients considerable time, effort, and money.

    There are many unanswered questions (for both lecanemab and donanemab), including the need for indefinite treatment and the timing of restarting treatment or maintenance dosing. We will need biomarker and clinical data to answer these over the next few years.

    While it is difficult to compare directly between clinical trials, it appears that ARIA rates may be a bit higher with donanemab. Efficacy may also be slightly better on primary outcome measures. Thus, it is possible that lecanemab may be a better choice for higher-risk patients, such as ApoE4 homozygotes, though real data will be needed to conclude this. Patient screening criteria and the infrastructure and expertise needed to administer donanemab is likely to be the same as that needed for lecanemab, and monitoring for adverse events should be similar, so I expect that integrating donanemab into practice will be straightforward. Ultimately, issues like cost and insurance coverage will play a major role.

    I expect that patients desire donanemab over lecanemab in most cases due to the convenience factor, though the possibility of lower ARIA risk will push some toward lecanemab.

  6. The use of a baseline tau threshold was an innovative feature of the donanemab trial, and this decision led to the discovery that baseline tau likely influences both treatment-related amyloid clearance and cognitive benefit, with the highest-tau individuals showing less cognitive benefit, perhaps due to more progressed disease. A patient's degree of tau burden likely has further implications for treatment that we do not yet fully understand. The "very low tau/no tau" group in particular is the group we know the least about. Although Lilly demonstrated that there is amyloid clearance in this group, there may be characteristics of this group that differ from the higher-tau groups enrolled in the trial, and this could affect the generalizability of treatment.

    In a recent analysis of "low tau" individuals on the AD pathway (amyloid-positive, impaired) from the Alzheimer's Disease Neuroimaging Initiative (Landau et al., 2024), we found that "low tau" is not rare; about 20-40 percent of amyloid-positive, impaired individuals had tau PET burden throughout cortex that is within the normal range. These low-tau patients were more likely to be male, older, and have more cardiovascular risk factors compared to their "higher tau" counterparts. Also, cognition was disproportionately low in those who had low tau burden together with α-synuclein burden. These findings suggest that people in the low-tau group may have other comorbid conditions that affect cognition (vascular disease, Lewy Body Dementia) that could reduce the cognitive benefits of amyloid-modifying therapy.

    The lack of donanemab treatment influence on tau outcomes adds further concern and confusion to the interpretation of treatment-related cognitive benefits. (I would note that there is no consensus on how to define low tau, so the group we identified may be different from the low-tau participants in the donanemab trial. Also, tau PET tracers are not sensitive to early stage tau, so even those in the low-tau group may have substantial burden.)  

    There are many unanswered questions but the low-tau PET group is probably relatively common in the patient population, and these patients may have key demographic and comorbid characteristics that differ from higher-tau individuals that could influence how treatment plays out in the clinic.


    . Individuals with Alzheimer's disease and low tau burden: Characteristics and implications. Alzheimers Dement. 2024 Mar;20(3):2113-2127. Epub 2024 Jan 19 PubMed.

  7. I was unable to tune in for the hearing, but from what I have heard from colleagues, we are overall happy to see the committee’s approval. I understand the committee’s reasoning in not requiring tau PET. I also agree with Dr. Knopman’s point in his Alzforum commentary that it would be helpful at specialty centers and could help accrue real-world data on how well these medications work in people with different levels of tau accumulation.

    I agree the rates of ARIA-E, especially symptomatic ARIA-E, in apoE4 homozygotes are a concern. Overall, donanemab seems to have slightly higher ARIA rates than lecanemab and similar efficacy, so not clearly a better drug efficacy versus side effects. But the monthly dosing will be attractive to patients and families and will put less burden on already-strained infusion centers. 

    Once donanemab becomes available at our center, I anticipate we will have a discussion of these pros and cons with patients and families based on their risk profiles. It might be reasonable to prefer lecanemab in ApoE4 homozygotes, for example, or for those who are more risk-averse. Patients who don't like to travel and/or live a long way from an infusion center might prefer donanemab. 

    As with lecanemab, it will be critically important to collect “real world” data and outcomes to better understand how these drugs work over time and to begin to answer the outstanding questions, like how long to treat. The best thing about this approval is that it is hopefully a sign that our pharma partners will remain committed to AD research and that we will have more trials, and more effective medications, in the next few years.

Make a Comment

To make a comment you must login or register.


Therapeutics Citations

  1. Donanemab
  2. Leqembi
  3. Aduhelm

News Citations

  1. Unanimous: FDA Advisory Committee Backs Donanemab
  2. Donanemab Data Anchors Upbeat AAIC
  3. And Then There Were Three: Donanemab Phase 3 Trial Positive
  4. Donanemab Confirms: Clearing Plaques Slows Decline—By a Bit
  5. Gaining a Foothold: Amyloid Immunotherapy in Clinical Practice
  6. Can Donanemab Prevent AD? Phase 3 TRAILBLAZER-ALZ3 Aims to Find Out
  7. Donanemab Phase 3 Puts Plasma p-Tau, Remote Assessments to the Test
  8. CSF MTBR-tau-243 Tracks Tangles, Plummets in Response to Antibody
  9. Tau Fragments in Plasma Track with Tangles, Cognitive Decline
  10. Is ARIA an Inflammatory Reaction to Vascular Amyloid?

Paper Citations

  1. . Aducanumab anti-amyloid immunotherapy induces sustained microglial and immune alterations. J Exp Med. 2024 Feb 5;221(2) Epub 2024 Jan 16 PubMed.

Further Reading