Lecanemab and donanemab: state of practice for anti-amyloid mAbs

Why these drugs matter

For 20 years, Alzheimer's treatment meant symptomatic drugs: donepezil, rivastigmine, galantamine, memantine. All four modestly improve cognitive symptoms and delay institutionalization by weeks, on average. None of them touch the underlying disease. Every anti-amyloid trial in the 2000s and early 2010s failed, some spectacularly (bapineuzumab, solanezumab, gantenerumab).

Aducanumab (Aduhelm) received an accelerated approval in 2021 that was widely criticized because the Phase 3 program had mixed results and the FDA advisory committee had voted against approval. Aducanumab was discontinued in 2024.

Lecanemab and donanemab are different. Both had positive Phase 3 trials with pre-specified primary endpoints that were met. Both showed reduced amyloid burden on PET scans that correlated with clinical benefit. Both received full traditional FDA approval rather than accelerated approval. This is the first genuine disease-modifying therapy for Alzheimer's, meaning treatment that changes the trajectory of the disease rather than just easing symptoms.

The word "modification" needs precision, though. Both drugs slow decline. Neither reverses it. Neither restores lost cognitive function. Patients starting these drugs continue to worsen, just more slowly than they would have. That's a real advance, but framing it accurately for patients and families is essential to informed consent.

What the trials showed

Lecanemab: CLARITY-AD (van Dyck et al., NEJM 2023), 1,795 patients with early AD, 18 months of biweekly infusions. Primary endpoint was change on the Clinical Dementia Rating Sum of Boxes (CDR-SB). Lecanemab slowed decline by 0.45 points on CDR-SB, roughly a 27 percent slowing versus placebo. Secondary endpoints (ADCS-ADL, ADAS-Cog14) all favored lecanemab. Amyloid PET showed near-complete plaque removal by month 18 in most patients. ARIA-E occurred in 12.6 percent of the lecanemab group versus 1.7 percent of placebo; ARIA-H in 17.3 percent versus 9 percent.

Donanemab: TRAILBLAZER-ALZ 2 (Sims et al., JAMA 2023), 1,736 patients with early AD, 72 weeks of monthly infusions. Primary endpoint was integrated Alzheimer's Disease Rating Scale (iADRS). Donanemab slowed decline by about 22 percent overall, 35 percent in low-medium tau subgroup. Amyloid PET clearance was substantial and often allowed treatment discontinuation. ARIA-E occurred in 24 percent (higher than lecanemab); ARIA-H in 31.4 percent. Three deaths in the trial were plausibly related to ARIA.

Both trials excluded patients on anticoagulation, patients with prior cerebral hemorrhage, patients with more than 4 microhemorrhages on baseline MRI, and patients with significant cerebral amyloid angiopathy findings. Exclusions that patients meet in real practice are common, and the label restrictions largely mirror the trial exclusions.

ARIA: what it is and how it's monitored

Amyloid-related imaging abnormalities (ARIA) are the class-defining safety concern. Two subtypes:

ARIA-E is edema, meaning fluid accumulation in the brain parenchyma or meninges, seen as increased signal on FLAIR MRI. Most cases are asymptomatic and detected only on scheduled surveillance MRI. Symptomatic cases (headache, confusion, focal deficits, seizure) are less common but do happen.

ARIA-H is hemorrhage, meaning cerebral microhemorrhages or superficial siderosis, seen as low signal on gradient echo or susceptibility-weighted MRI. Usually asymptomatic. Large parenchymal hemorrhages are rare but have been reported and have caused deaths.

Both drugs require baseline MRI within 1 year before starting. Lecanemab requires surveillance MRIs before infusion 5, 7, and 14 (that's approximately months 3, 6, and 12 in the biweekly schedule). Donanemab requires MRIs before infusion 2, 3, 4, and 7 (months 1, 2, 3, and 6). Any patient with new symptoms suggesting neurologic change gets an unscheduled MRI.

ARIA management depends on severity. Mild asymptomatic ARIA-E allows continued dosing with next scheduled MRI check. Moderate to severe ARIA-E, or symptomatic ARIA of any grade, means pausing infusions and rescanning until resolution. Persistent or severe ARIA usually means permanent discontinuation. Detailed management algorithms are in the labels and in the Alzheimer's Association / American Academy of Neurology appropriate use recommendations.

ApoE4 genotype matters here. ApoE4 homozygotes have roughly triple the ARIA risk of ApoE4 non-carriers. Both labels recommend disclosing ApoE4 status to patients before starting treatment and factoring it into shared decision-making. Some clinicians will not treat ApoE4 homozygotes given the elevated risk; others will treat with enhanced monitoring after explicit informed consent. There's no universal rule, but the conversation is not optional.

Patient selection

The label indications for both drugs are essentially identical: mild cognitive impairment or mild dementia due to Alzheimer's disease, with confirmed amyloid pathology. The Alzheimer's Association appropriate use recommendations added further clarification, and in practice the criteria are:

Cognitive: MMSE 22 to 30 (some centers use MoCA 17 to 30). Clinical Dementia Rating (CDR) global score of 0.5 (MCI) or 1 (mild dementia). Patient is aware of memory problems and capable of consent (with a caregiver participating).

Confirmation of amyloid pathology: positive amyloid PET scan, or CSF markers (Ab42/40 ratio or Ab42/tau ratio) consistent with AD. Blood-based biomarkers (plasma p-tau217, Precivity AD2) are being adopted but are not yet universally accepted for treatment eligibility in the way PET or CSF is. This is changing quickly.

Exclusions (both drugs): significant vascular dementia component (not pure AD), current anticoagulation with warfarin, DOAC, or heparin, more than 4 microhemorrhages on baseline MRI, prior cerebral macrohemorrhage, significant cerebral amyloid angiopathy findings, uncontrolled hypertension, active malignancy, life expectancy less than 2 years.

The exclusion for anticoagulation is a real barrier. Many patients with AD have atrial fibrillation and are on DOACs. Weighing stroke risk from stopping the anticoagulant against ARIA-H risk from continuing it is a complex individualized decision, and neither trial included these patients.

Infusion logistics

Lecanemab is 10 mg/kg IV every two weeks. Infusion time is 1 hour after standard hydration and premedication. The biweekly schedule means 26 infusions per year, and this is a significant time commitment for the patient and caregiver. Home infusion is starting to be available in some regions.

Donanemab is 700 mg IV every four weeks for the first three doses, then 1400 mg every four weeks. Infusion time is 30 minutes. Monthly schedule is more manageable for many families.

A key difference: donanemab treatment can potentially be stopped once amyloid PET shows plaque clearance (typically 12 to 18 months). The trial protocol allowed discontinuation once amyloid PET fell below 24 centiloids, and roughly half of donanemab-treated patients met this threshold by 12 months. Lecanemab's protocol did not include a stopping rule and current practice is generally to continue indefinitely, though this is under active discussion.

The stopping question matters clinically and economically. If donanemab produces durable amyloid clearance and cognitive benefit persists after stopping, the total treatment burden and cost are lower than lecanemab's indefinite biweekly infusion.

Cost, coverage, and CMS requirements

Lecanemab's annual cost is roughly $26,500. Donanemab's is about $32,000. Additional costs (amyloid PET, MRI monitoring, ApoE4 genotyping, infusion administration) add several thousand dollars per year.

CMS coverage for both drugs is conditional on participation in an approved registry. That registry captures real-world data on efficacy, safety, and patient characteristics. Medicare Part B covers the drug and administration in approved settings when the coverage requirements are met. Patients typically pay a 20 percent coinsurance unless they have supplemental insurance.

Practical implication: most patients need to be treated at a center with the infrastructure for amyloid PET or CSF diagnosis, ApoE4 genotyping, MRI monitoring on the required schedule, and infusion services with the registry submission. This is not primary care territory. Referral to a memory clinic or neurology practice is usually required.

Where anti-amyloid mAbs fit right now

Three patient groups where the case is clearest in mid-2026:

The first is patients with confirmed amyloid-positive early AD (MCI or mild dementia), no significant vascular disease, no anticoagulation, no or few microhemorrhages, ApoE4 non-carrier or heterozygote, engaged patient and family who understand the modification-not-cure framing, and reasonable access to the infusion infrastructure. This patient is the trial patient and is the strongest indication.

The second is patients where family history of AD and personal cognitive change are motivating pre-symptomatic testing, and the amyloid workup confirms disease. Here anti-amyloid therapy at the earliest end of the disease spectrum has the best-hypothesized benefit. The Phase 3 trials didn't specifically enroll pre-symptomatic patients, so this is extrapolation, but the pathophysiology supports treating earlier rather than later. Ongoing prevention trials (AHEAD-3, TRAILBLAZER-ALZ 3) are testing this directly.

The third is patients with moderate dementia who are still functionally engaged and where family wants to try slowing further decline. Off-label in the sense that the label is "mild," but some practices treat MCI-to-moderate spectrum patients when the clinical picture supports it. Effect sizes at more advanced disease stages are not established.

Patient groups where these drugs are not first-line: significant vascular contribution to cognitive symptoms, current anticoagulation with stroke history, ApoE4 homozygotes without extensive counseling, patients who cannot commit to MRI monitoring or infusion schedule, patients with life expectancy under 2 years, patients with non-AD dementia (Lewy body, frontotemporal, vascular).

Common questions

Do these drugs work? Yes, in the specific sense that they slow cognitive decline. Both trials met their primary endpoints with clinically meaningful effect sizes. They do not stop the disease, do not reverse it, and do not restore lost function. The framing is a slower trajectory, not a cure.

Which drug is better? No head-to-head trial exists. Cross-trial comparisons are hazardous but the effect sizes look broadly similar. Donanemab's ARIA rate is higher; lecanemab's infusion burden is higher. Donanemab may allow treatment discontinuation once amyloid clears; lecanemab is currently indefinite. Choice often comes down to logistics (biweekly IV clinic access), risk tolerance (ApoE4 status, MRI comfort), and local availability.

How long does treatment continue? Lecanemab: currently indefinite based on trial protocol and current practice. Whether long-term treatment is necessary once amyloid clears is under study.

Donanemab: treatment can be stopped when amyloid PET shows clearance below the pre-specified threshold. Roughly 50 percent of trial patients reached this by 12 months and 76 percent by 18 months. Whether cognitive benefit persists after stopping is being tracked; interim data suggests it does for at least a year.

What if a patient has an ARIA event? Depends on grade and symptoms. Mild asymptomatic ARIA usually allows continued dosing with more frequent MRI. Moderate or symptomatic ARIA usually means pausing infusions and rescanning until resolution. Severe ARIA (large hemorrhage, extensive edema, symptomatic seizure) usually means permanent discontinuation. Formal grading is in both product labels and in the AA/AAN appropriate use recommendations.

Should ApoE4 genotyping be done before treatment? Yes, both labels recommend it. ApoE4 homozygotes have substantially higher ARIA risk and the informed consent discussion changes materially. Some centers will not treat ApoE4 homozygotes; others will, with enhanced monitoring. The genotype should be disclosed to the patient with genetic counseling in most cases, since it affects Alzheimer's risk stratification for the patient and potentially for family members.

Can these drugs be combined with cholinesterase inhibitors or memantine? Yes, and typically are. Symptomatic treatment (donepezil, memantine) continues in parallel with anti-amyloid therapy. There's no known negative interaction. The trials allowed concomitant symptomatic AD drugs, and most trial patients were on them.

What about vascular risk factors? Hypertension, atrial fibrillation, and prior stroke change the ARIA risk-benefit calculation and often make anti-amyloid therapy inappropriate. Pre-treatment cardiovascular optimization matters both for baseline safety and for the general dementia trajectory. Uncontrolled hypertension is often a treatment exclusion.

Is amyloid PET or CSF required? Yes for both drugs. Either amyloid PET (Amyvid, Vizamyl, Neuraceq) or CSF (Ab42/40 or Ab42/p-tau ratio) is required to confirm amyloid pathology. Blood-based biomarkers (plasma p-tau217, Precivity AD2) are becoming more accepted and may replace CSF and PET as the primary confirmation method within the next few years. Current practice is that if a positive blood test is followed by a confirmatory PET or CSF, the confirmatory test is what documents eligibility for the label.

Can patients be treated in primary care or in psychiatry? The label doesn't restrict prescribing to neurology, but the practical infrastructure requirements (amyloid confirmation, MRI monitoring, infusion services, ARIA management, registry participation) usually mean referral to a memory clinic or neurology practice. Psychiatry's role in this population is often the concurrent management of depression, agitation, sleep, and behavioral symptoms while the neurology team runs the anti-amyloid protocol.

Does treatment prevent progression to more severe dementia? No trial has followed patients long enough to establish that. The 18-month CLARITY-AD data and 72-week TRAILBLAZER-ALZ 2 data both show slowed decline. Extension data suggests continued benefit but not prevention. The pre-symptomatic prevention trials (AHEAD-3, TRAILBLAZER-ALZ 3) are testing whether treating amyloid-positive but cognitively normal patients delays or prevents onset. Results are expected in 2027 to 2029.

What's next

The class is expanding. Remternetug (Eli Lilly), an anti-amyloid antibody with subcutaneous dosing, is in late-stage trials. If it works, self-administered subcutaneous injection would remove the infusion burden and lower the total cost of care. Trontinemab (Roche) is a brain-shuttle antibody that may cross the blood-brain barrier more efficiently, potentially reducing ARIA.

The bigger picture question is whether combination therapy targeting amyloid AND tau produces larger effect sizes. Anti-tau immunotherapy trials are ongoing. If the two mechanisms combine additively, the modification-of-decline signal may become large enough to talk about arrest or reversal rather than slowing.

The pre-symptomatic prevention question is the most important open one. If treating amyloid-positive but cognitively normal patients prevents dementia, the drugs shift from disease-modifying therapy to prevention, which changes health system planning fundamentally. If it doesn't, the drugs remain a slowing therapy for a narrow slice of the AD population.

Sources

• van Dyck CH, Swanson CJ, Aisen P, et al. Lecanemab in early Alzheimer's disease. N Engl J Med. 2023;388(1):9-21. (CLARITY-AD)
• Sims JR, Zimmer JA, Evans CD, et al. Donanemab in early symptomatic Alzheimer's disease: the TRAILBLAZER-ALZ 2 randomized clinical trial. JAMA. 2023;330(6):512-527.
• Leqembi (lecanemab-irmb) injection. Prescribing information. Eisai and Biogen; traditional approval July 6, 2023.
• Kisunla (donanemab-azbt) injection. Prescribing information. Eli Lilly; approved July 2, 2024.
• Cummings J, Apostolova L, Rabinovici GD, et al. Lecanemab: appropriate use recommendations. J Prev Alzheimers Dis. 2023;10(3):362-377.
• Rabinovici GD, Selkoe DJ, Schindler SE, et al. Donanemab: appropriate use recommendations. J Prev Alzheimers Dis. 2025.
• Centers for Medicare and Medicaid Services. National Coverage Determination for Monoclonal Antibodies Directed Against Amyloid for the Treatment of Alzheimer's Disease.