Alzheimer Disease in 2026: Blood-Biomarker Diagnosis, Anti-Amyloid Therapy, and the Care Pathway

Bottom Line:

• Alzheimer disease (AD) is now defined biologically in the 2024 Alzheimer's Association Workgroup criteria; a validated, sufficiently accurate "Core 1" biomarker — CSF Aβ42/40, CSF p-tau181/Aβ42, or a comparably accurate plasma p-tau217/p-tau181/p-tau231 assay — can establish AD in a symptomatic patient (Jack, Alzheimers Dement 2024).
• The FDA has now cleared multiple plasma AD diagnostic aids for symptomatic adults being evaluated in specialized care: the Fujirebio Lumipulse G pTau217/β-Amyloid 1-42 Plasma Ratio (May 2025), Roche Elecsys pTau181 plasma (Oct 2025), PrecivityAD2 mass-spectrometry plasma (K253240, Aug 19 2026), and Roche Elecsys pTau217 plasma (K261686, Aug 2026). Each product's cleared indication is symptomatic evaluation, not screening. An open FDA Class II recall (2026) covers specified Lumipulse ratio lots that produced falsely elevated results — verify lot numbers and revisit prior classifications.
• Two anti-amyloid monoclonals have traditional FDA approval: lecanemab (CLARITY AD, NEJM 2023) and donanemab (TRAILBLAZER-ALZ 2, JAMA 2023). Absolute clinical effects are modest and, for donanemab, larger in the low/medium-tau subpopulation. The current Kisunla USPI prescribes a slower titration — 350 → 700 → 1050 → 1400 mg IV every 4 weeks — added to the label in 2025 on the strength of TRAILBLAZER-ALZ 6; the 700 mg × 3 then 1400 mg schedule is the original TRAILBLAZER-ALZ 2 study regimen, not current prescribing.
• The current Leqembi USPI specifies MRIs "after 1 month, 2 months, 3 months, and 6 months of treatment"; the Kisunla USPI uses a different, infusion-anchored schedule (prior to the 2nd, 3rd, 4th, and 7th infusions). Do not conflate the two.
• In July 2026 the FDA approved Leqembi IQLIK 500 mg SC weekly as starting therapy; the label requires initiation under a healthcare provider's guidance and supervision for at least two consecutive SC doses before patient or caregiver home administration may be considered. After 18 months the label permits continuation of the starting regimen or transition to 10 mg/kg IV every 4 weeks or 360 mg SC weekly maintenance. Approval rests on PK bridging and IV pivotal outcomes, not an independent SC clinical-outcome trial.
• APOE ε4 genotyping is recommended prior to initiating either monoclonal to inform ARIA risk. Neither label excludes APOE ε4 homozygotes and neither prescribes a different MRI cadence for them.
• Both Appropriate Use Recommendations advise against anti-amyloid therapy in patients on chronic anticoagulation; the donanemab AUR states anticoagulation should not be stopped merely to enable treatment. The Leqembi USPI also carries an explicit Warning against concomitant thrombolytic (tPA) use because of fatal intracerebral hemorrhage — this is a label-level warning, not just AUR guidance.
• Oral semaglutide 14 mg failed on all clinical endpoints in the twin EVOKE / EVOKE+ phase-3 trials in early AD (Lancet 2026) despite CSF biomarker movement — downstream biomarker signal does not equal clinical benefit.

The basics

Definition

The 2024 revised Alzheimer's Association Workgroup criteria (Jack et al., Alzheimers Dement 2024) reframe AD as a biological diagnosis. Biomarkers are grouped as Core 1 — early-changing markers of Aβ pathology and early tau (CSF Aβ42/40, CSF p-tau181/Aβ42, and plasma p-tau217, p-tau181, p-tau231 when a sufficiently accurate assay is used); Core 2 — later-changing markers, including tau PET, MTBR-tau243, pT205, and non-phosphorylated mid-region tau fragments; and non-specific markers of neurodegeneration and inflammation. The Workgroup states that AD can be diagnosed in a symptomatic patient by an abnormal validated, sufficiently accurate Core 1 assay, and cautions against clinical use in asymptomatic individuals outside research. Establishing biological AD under the 2024 criteria is not the same as meeting the amyloid-pathology confirmation the anti-amyloid AURs require before disease-modifying therapy — that decision currently uses PET or CSF (see §"Care pathway").

Epidemiology

The Alzheimer's Association 2026 Facts and Figures report estimates roughly 7 million US adults age ≥65 living with AD dementia, with age-specific incidence rising steeply after 75 and persistent racial/ethnic disparities in diagnostic latency.

Pathophysiology in one paragraph

Extracellular Aβ plaques and intraneuronal hyperphosphorylated tau (tangles) are the defining lesions. Aβ deposition typically precedes symptomatic decline by 15–20 years; downstream tau spread from medial temporal lobe to neocortex tracks more closely with the clinical trajectory. Microglial and complement-mediated synaptic pruning, astrogliosis (GFAP), and axonal injury (NfL) complete the picture. APOE ε4 reduces Aβ clearance in a dose-dependent way and increases risk of both AD and cerebral amyloid angiopathy (CAA).

Phenotypes

Amnestic AD accounts for approximately 85% of cases. Atypical variants — posterior cortical atrophy (biparietal), logopenic primary progressive aphasia, and dysexecutive/behavioral AD — share the AD biomarker signature but present with non-memory-dominant deficits (Graff-Radford et al., Lancet Neurol 2021).

How AD was diagnosed historically

Through 2023, AD in community practice rested on NIA-AA 2011 clinical criteria; CSF Aβ42/40 with p-tau181 and amyloid PET were confined to specialty and research settings.

Where we are

Symptomatic pharmacotherapy — one compact paragraph

Cholinesterase inhibitors (donepezil, rivastigmine, galantamine) provide modest cognitive/functional benefit in mild-to-moderate AD, class-wide, without a head-to-head winner. Memantine adds small benefit in moderate-to-severe disease. This class is unchanged from 2010.

Agitation

Brexpiprazole is the only agent FDA-approved for agitation associated with dementia due to AD (approval May 11, 2023). In the pivotal 12-week trial (Lee et al., JAMA Neurol 2023;80:1307–1316), brexpiprazole 2 or 3 mg/d reduced Cohen-Mansfield Agitation Inventory total score more than placebo (LSM difference −5.32; 95% CI −8.77 to −1.87; P=0.003; Cohen d 0.35). Brexpiprazole, like all atypical antipsychotics, carries a Boxed Warning for increased mortality in elderly patients with dementia-related psychosis; this label warning applies to brexpiprazole use in this indication and must be disclosed to patients and families. Pimavanserin (HARMONY, NEJM 2021;385:309–319) addresses a different indication — dementia-related psychosis, not AD-associated agitation — and is not FDA-approved for dementia-related psychosis; do not conflate the two agents.

Non-pharmacologic

The MIND diet observational signal (Morris et al., Alzheimers Dement 2015; β=0.0092 per year, P<0.0001; highest vs lowest tertile equivalent to being 7.5 years younger) is hypothesis-generating. The subsequent randomized MIND-diet trial (Barnes et al., NEJM 2023) found no significant cognitive or MRI advantage over a mild-calorie-restriction control at 3 years — the observational signal has not translated into randomized benefit. Sleep, exercise, vascular-risk control, and hearing correction remain standard advice.

Historical amyloid-antibody attempts — the scaffold, not the story

Bapineuzumab 301/302 (NEJM 2014) were negative. Solanezumab (EXPEDITION-1, 2014; A4 in preclinical AD, NEJM 2023;389:1096–1107) showed no clinical benefit. Aducanumab (EMERGE positive high-dose CDR-SB −0.39, P=0.012; ENGAGE negative; pooled analysis J Prev Alz Dis 2022) received accelerated approval in 2021 and was withdrawn by the sponsor in 2024. These are the scaffold for what changed in 2023–2026 and are not restated below.

What is new

Blood-biomarker diagnosis: three ideas the reader must keep separate

Three distinct claims travel together in the press coverage and are frequently merged in clinic. Keep them apart.

1. What the 2024 criteria permit. Under Jack et al. 2024, a validated, sufficiently accurate Core 1 plasma assay (p-tau217, p-tau181, or p-tau231) can establish biological AD in a symptomatic patient without CSF or PET confirmation. This is a criteria-level allowance, not a treatment threshold.

2. What the FDA has cleared. The regulatory landscape moved rapidly between May 2025 and August 2026. Four blood-based diagnostic aids now hold FDA clearance for symptomatic adults being evaluated for AD in specialized care: the Fujirebio Lumipulse G pTau217/β-Amyloid 1-42 Plasma Ratio (K242706, May 2025); Roche Elecsys pTau181 plasma (K252163, October 2025); C2N PrecivityAD2, a mass-spectrometry plasma test (K253240, August 19 2026); and Roche Elecsys pTau217 plasma (K261686, August 2026). Each clearance is a device-specific indication for use as an aid in evaluating symptomatic patients being evaluated for AD; none is cleared for asymptomatic screening. In the FDA-reviewed Lumipulse ratio dataset, 91.7% of positive results and 97.3% of negative results agreed with amyloid pathology, where the reference standard combined amyloid PET or an FDA-authorized CSF test. A Class II recall / product correction (2026, Res ID 217945) covers specified lots of the Lumipulse ratio because falsely elevated ratios produced excess indeterminate or positive classifications and lower specificity — laboratories should verify lot numbers and, where applicable, reassess prior classifications.

3. What anti-amyloid treatment workflows require. Both the lecanemab AUR (Cummings 2023) and the donanemab AUR (Rabinovici 2025) require confirmed amyloid pathology before initiating disease-modifying therapy. In practice, both AURs treat plasma assays as high-value triage but confirm pathology with CSF or amyloid PET before treatment. This may relax as clearances accumulate and post-marketing data mature, but is the current AUR posture.

The validation science (2025). Palmqvist et al. (Nat Med 2025;31:2036–2043) evaluated the fully automated Lumipulse plasma p-tau217 immunoassay (standalone p-tau217) across five cohorts. Single-cutoff accuracy was 89–91% in secondary care and 85% in primary care; a two-cutoff strategy (0.34 pg/mL) reached 92–94% accuracy with 12–17% of samples falling intermediate. Performance was preserved across sex, APOE genotype, diabetes, and CKD; accuracy in patients ≥80 y was 83% with a single cutoff (vs 91% overall, P=0.003), rescued by the two-cutoff approach. This standalone p-tau217 assay and the FDA-cleared ratio use different analytes and different cutoffs — do not carry one paper's cut-points across products.

A misconception to close down. Plasma p-tau217 cannot be used at the individual level to confirm treatment-related amyloid clearance after donanemab. In the TRAILBLAZER-ALZ-2 p-tau217 substudy (Collins et al., Alzheimers Dement 2026;22:e71740; N=830 study population, week-52 mass-spectrometry subset N=670), the mass-spectrometry p-tau217 AUROC for detecting <24.1 CL clearance was 0.61 at 52 wk (0.64 at 24 wk; 0.63 at 76 wk); the Elecsys immunoassay performed similarly. A multivariate model (p-tau217 + p-tau181 + GFAP + NfL + age + APOE ε4) reached only AUROC 0.71 at 76 wk. Amyloid PET remains necessary to confirm plaque clearance.

Trajectory-based approaches ("p-tau217 clocks," Nat Med 2026) and use of plasma p-tau217 as a trial endpoint (Ferreira et al., Neurology 2026) are being developed but are not for clinical decision-making today.

Anti-amyloid antibodies in practice: lecanemab and donanemab

Lecanemab (Leqembi). CLARITY AD (van Dyck et al., NEJM 2023;388:9–21) randomized 1,795 participants with early symptomatic AD (MCI or mild dementia, biomarker-confirmed, MMSE 22–30) to lecanemab 10 mg/kg IV every 2 weeks vs placebo for 18 months. Primary endpoint: CDR-SB change at 18 mo was 1.21 (lecanemab) vs 1.66 (placebo) (adjusted MD −0.45; 95% CI −0.67 to −0.23; P<0.001) — 27% relative slowing. Key secondaries (ADAS-cog14, ADCOMS, ADCS-MCI-ADL) all favored lecanemab (P<0.001). Amyloid PET reached a mean of 22.99 CL. ARIA-E: 12.6% overall (2.8% symptomatic), 32.6% in APOE ε4 homozygotes vs 10.9% heterozygotes vs 5.4% non-carriers; symptomatic ARIA-E 9.2% / 1.7% / 1.4%. Infusion-related reactions 26.4%. Traditional FDA approval July 2023.

Subcutaneous lecanemab (IQLIK). The 360 mg SC once-weekly maintenance formulation was approved in 2025 (as a switch after 18 months of IV). In July 2026, the FDA approved 500 mg SC once weekly as starting therapy. The current Leqembi USPI reports approximately 77.5% absolute bioavailability for the 500 mg SC autoinjector regimen and 53% for the 360 mg SC maintenance regimen; the earlier healthy-volunteer PK work (Penner et al., Alzheimers Dement 2025) reported 49.7% (90% CI 43.5–56.8) for an earlier SC vial formulation and is a scientific reference — not the label PK. Efficacy for SC starting therapy was extrapolated from IV CLARITY AD via exposure-response modeling, supported by safety data across 424 SC recipients (72 treatment-naïve) in the current label; there is no independent SC clinical-outcome trial. The label requires that SC therapy be initiated under the healthcare provider's guidance and supervision, and permits patient or caregiver home administration only after direct guidance during at least two consecutive SC doses and a provider determination that home administration is appropriate. After 18 months, the label permits continuation of the starting regimen, transition to 10 mg/kg IV every 4 weeks, or transition to 360 mg SC weekly, including transitions from SC starting therapy.

Donanemab (Kisunla). TRAILBLAZER-ALZ 2 (Sims et al., JAMA 2023;330:512–527) enrolled 1,736 participants with early symptomatic AD (MMSE 20–28), prestratified by tau PET into low/medium and high tau. In the low/medium-tau coprimary population, iADRS change at 76 wk was −6.02 vs −9.27, a 3.25-point difference (95% CI 1.88 to 4.62; P<0.001) — 35.1% slowing. CDR-SB change was 1.20 vs 1.88 (P<0.001). In the combined coprimary population, iADRS change was −10.19 vs −13.11 (P<0.001) — approximately 22.3% slowing; CDR-SB change 1.72 vs 2.42 (P<0.001). Amyloid clearance (<24.1 CL) at 76 wk was 80.1% in low/medium-tau and 76.4% in the combined population. ARIA-E occurred in 24% of donanemab recipients (6.1% symptomatic); 3 treatment-related deaths were attributed to serious ARIA. Traditional FDA approval July 2024. When counseling patients, report both populations — do not quote the low/medium-tau numbers alone.

The current Kisunla titration is not the TRAILBLAZER-ALZ 2 titration. TRAILBLAZER-ALZ 6 (Wang et al., Alzheimers Dement 2025;21:e70062) tested a slower up-titration (350 → 700 → 1050 → 1400 mg IV every 4 weeks). ARIA-E at 24 wk was 13.7% (modified) vs 23.7% (standard) — posterior RRR 0.405; 94.1% posterior probability of ≥20% RRR; through 52 wk, 15.6% vs 24.2%. Amyloid clearance (58.8 vs 56.3 CL adjusted mean reduction) and p-tau217 lowering were equivalent. The FDA added this schedule to the Kisunla label in 2025; the current USPI initiation schedule is 350 → 700 → 1050 → 1400 mg IV every 4 weeks. The original 700 mg × 3 then 1400 mg regimen from TRAILBLAZER-ALZ 2 is historical trial dosing and not current prescribing.

Who to consider treating — and who not to

The two Appropriate Use Recommendations — Cummings et al. for lecanemab (J Prev Alz Dis 2023) and Rabinovici et al. for donanemab (J Prev Alz Dis 2025) — mirror pivotal-trial enrollment: MCI or mild dementia due to AD, biomarker-confirmed, with the MMSE ranges used in the pivotal trials (22–30 in CLARITY AD; 20–28 in TRAILBLAZER-ALZ 2). Neither the donanemab USPI nor its AUR requires tau PET for eligibility; the donanemab AUR states tau PET is not required and may be used, where available, to individualize expected benefit. Eligibility, exclusion, label, and AUR guidance are not identical and should not be merged. Table 4 preserves the source of each row.

On anticoagulation and thrombolytics. Chronic anticoagulation is not a labeled contraindication in either USPI, but both AURs advise against treatment; in practice, patients on apixaban, warfarin, or another chronic anticoagulant are excluded from anti-amyloid therapy by AUR guidance. For lecanemab specifically, the Leqembi USPI Warnings and Precautions explicitly warn of fatal intracerebral hemorrhage with concomitant tPA — this is an FDA label warning, and emergency clinicians should treat it as such when a lecanemab-treated patient presents with acute stroke syndromes.

APOE ε4 stratification

Both labels recommend APOE ε4 genotyping prior to initiation to inform ARIA risk. In CLARITY AD the ARIA-E gradient was 5.4% (non-carrier) → 10.9% (heterozygote) → 32.6% (homozygote); symptomatic ARIA-E was 1.4% / 1.7% / 9.2%. The labels do not exclude APOE ε4 homozygotes and do not prescribe a different imaging cadence. Homozygotes warrant an explicit conversation about magnitude of benefit, elevated symptomatic-ARIA risk, and the option to decline; genetic counseling for the patient and biological relatives should be offered before the swab, not after. Dose-modification concepts in homozygotes (e.g., K-ROAD registry data) are investigational and not label-based.

The care pathway — what a positive plasma result triggers

A positive plasma p-tau217 or plasma ratio in a symptomatic patient is a diagnostic aid that raises the pre-test probability of underlying AD pathology; under 2024 criteria it can establish biological AD, and under both AURs it currently prompts confirmatory pathology testing before disease-modifying therapy. A comprehensive care pathway includes at minimum:

1. History and cognitive/functional evaluation. Structured cognitive testing (MoCA or comparable), functional inventory (FAQ or CDR), collateral history, medication review (anticholinergic burden), and screening for depression, delirium, and OSA.
2. Differential diagnosis and co-pathology. DLB, FTD, PSP, vascular contribution, normal-pressure hydrocephalus, thyroid disease, B12 deficiency, and neurosyphilis where indicated. A positive amyloid biomarker does not rule out DLB or vascular contribution.
3. Basic labs and imaging. CBC, CMP, TSH, B12, HIV/RPR where appropriate; structural brain MRI is standard. Hemorrhage-sensitive sequences (SWI or T2*) are the practical way to characterize microhemorrhages, siderosis, and CAA features, per radiology/AUR guidance rather than a specific label mandate.
4. Staging. MCI vs mild dementia vs moderate/severe. Moderate-to-severe patients fall outside pivotal-trial evidence and labeled use for the two monoclonals.
5. Biomarker confirmation for treatment. Under the 2024 criteria a sufficiently accurate Core 1 plasma assay can establish biological AD; for anti-amyloid treatment eligibility, both AURs require confirmed amyloid pathology — typically CSF or amyloid PET.
6. APOE genotyping. Prior to initiation of a monoclonal, per each USPI, with pre-test counseling.
7. Shared decision-making. Magnitude of benefit (CDR-SB −0.45 in CLARITY AD; iADRS 3.25 in low/medium-tau donanemab, ~22.3% slowing in the combined population), ARIA risk stratified by APOE genotype, monitoring intensity, cost/coverage, driving/safety, caregiver burden.
8. Infrastructure. MRI access on the schedule the specific label prescribes; an infusion or supervised-then-home injection pathway; and a documented ARIA-emergency route. Community neurology without this infrastructure should refer.
9. Longitudinal reassessment. Cognitive/functional trajectory, safety, driving evaluation, advance care planning, and caregiver support. Kisunla labeling allows discontinuation upon documented amyloid clearance on PET; Leqembi does not prescribe a defined stop.

ARIA — quote the labels, do not paraphrase, and do not merge the two schedules

Lecanemab (Leqembi USPI, current). "Obtain a recent baseline brain magnetic resonance imaging (MRI) prior to initiating treatment with LEQEMBI. In addition, obtain MRIs after 1 month, 2 months, 3 months, and 6 months of treatment." The current label uses month-based timing and does not prescribe a separate schedule by APOE genotype.

Donanemab (Kisunla USPI). Baseline MRI plus MRIs prior to the 2nd, 3rd, 4th, and 7th infusions. This schedule is infusion-anchored and different from Leqembi's month-anchored schedule.

Do not invent a schedule the label does not prescribe. If a specific interval is not in the current USPI, do not write it. Check the current USPI before every prescribing decision — Leqembi's MRI language and IQLIK dosing were updated in 2025 and 2026, and the Kisunla titration was updated in 2025.

MRI sequence detail. The USPIs require baseline and surveillance MRI and define ARIA by FLAIR and microhemorrhage findings, but do not prescribe a complete acquisition protocol or explicitly mandate SWI/T2*. Recommendations for GRE/SWI, FLAIR, and DWI protocols come from the AURs and radiology guidance, not the labels themselves.

Clinical presentation of symptomatic ARIA. Headache, confusion, visual disturbance, gait instability, focal deficits, and — critically — seizure or status epilepticus. Both labels' Warnings and Precautions sections identify seizures explicitly. Suspected symptomatic ARIA warrants urgent brain MRI (using hemorrhage-sensitive sequences per institutional protocol) with dose modification per the specific label.

Cerebral amyloid angiopathy — the modifier under the hood

ARIA is mechanistically linked to CAA: anti-amyloid antibodies mobilize vascular Aβ, unmasking or exacerbating vessel-wall fragility. Pre-existing microhemorrhages, superficial siderosis, or a lobar microhemorrhage pattern consistent with probable CAA are the biologically rational trial-protocol exclusions and the reason the AURs are cautious about anticoagulation. Amyloid PET assesses cerebral plaque burden, not vascular fragility — microhemorrhages and siderosis are an MRI role, not a PET role. General frameworks: Greenberg et al., Nat Rev Neurol on CAA; Sperling et al., Lancet Neurol 2024 on ARIA pathophysiology.

What failed: GLP-1 in AD (EVOKE / EVOKE+)

EVOKE and EVOKE+ (Cummings et al., Lancet 2026;407:2167–2179) randomized 3,808 amyloid-confirmed participants with MCI or mild AD dementia across 566 sites in 40 countries to oral semaglutide titrated to 14 mg daily vs placebo for up to 156 weeks. The primary endpoint — CDR-SB change at week 104 — was not met in either trial: estimated difference −0.08 (95% CI −0.35 to 0.20; P=0.57) in EVOKE and 0.10 (95% CI −0.17 to 0.38; P=0.46) in EVOKE+. All confirmatory/supportive secondary endpoints (ADCS-ADL-MCI, ADAS-Cog-13, MoCA, MMSE, ADCOMS, NPI, composite Z-score, time to CDR-G ≥1.0) were null; the testing hierarchy was stopped after the primary failure. A CSF substudy (n=199) showed ~10% reductions in p-tau181, p-tau217, total tau, YKL-40, and neurogranin. Bodyweight fell 5.8% (−4.3 kg); TEAEs 91.2% vs 84.8%; discontinuation 17% vs 8%. Downstream biomarker movement is insufficient evidence to justify off-label use of GLP-1RAs for AD.

What is coming

AHEAD 3-45 (BAN2401-G000-303) — lecanemab in preclinical AD. A45: elevated amyloid (>40 CL), N~1,000, PACC5 primary. A3: intermediate amyloid (~20–40 CL), N~400, amyloid PET primary. 216-week treatment; uses C2N plasma Aβ42/40 for pre-screening (Rafii et al., Alzheimers Dement 2023).

TRAILBLAZER-ALZ 3 — donanemab in preclinical AD; in progress; anticipated read-out 2027.

DIAN-TU gantenerumab open-label extension (Bateman et al., Lancet Neurol 2025;24:316–330) — in dominantly inherited AD, 3-year PiB-PET SUVR change −0.71 (95% CI −0.88 to −0.53, P<0.0001); a substantial fraction of the longest-treated subgroup achieved amyloid-normal status. The longest-treated subgroup (~8.4 years) showed approximately 50% slowing of CDR-SB progression (HR 0.57, 95% CI 0.31–1.07) — sponsor-stopped, statistically uncertain, cite conservatively.

APOLLOE4 (valiltramiprosate) (Abushakra et al., Drugs 2025;85:1455–1472) — negative overall (ADAS-Cog13 slowing 11%, P=0.607), nominally significant in the prespecified MCI subgroup (52% slowing, nominal P=0.041); 18% hippocampal-atrophy slowing (P=0.017) overall; no ARIA. Hypothesis-generating for MCI-focused APOE ε4/ε4 studies; preserve the "nominally significant" hedge.

ABvac40 (AB1601) — 42-month safety extension of an active anti-Aβ40 vaccine (Pascual-Lucas et al., Alzheimers Dement 2026;22:e71746). No ARIA-E in either part; robust boostable anti-Aβ40 antibody response. The extension did not formally evaluate efficacy; exploratory clinical/imaging outcomes from Part A do not constitute confirmatory efficacy evidence.

ABBV-916 (Lynch et al., Alzheimers Dement 2026;22:e71715) — anti-Aβ pE3 monoclonal; dose-dependent amyloid lowering (−83.7 CL at 900 mg), overall ARIA-E 33.8% and ARIA-H 25.0%. Five of six APOE ε4 homozygotes experienced an ARIA event of any type in the reported cohort (the publication does not report this as ARIA-E specifically). Program discontinued for lack of differentiation.

Anti-tau combination. E2814 is under evaluation in the DIAN-TU platform in combination with anti-amyloid therapy in dominantly inherited AD.

Open questions

1. How long to treat? Leqembi does not define a firm stop; Kisunla allows discontinuation upon documented amyloid clearance on PET per the USPI. Whether re-treatment is required after PET reversion is unknown.
2. Is the individual clinical effect meaningful? −0.45 CDR-SB (CLARITY AD) and 3.25-point iADRS (TB-ALZ 2 low/medium-tau) sit below several proposed minimally important difference thresholds for MCI; patient-level meaningfulness remains contested.
3. Real-world ARIA and MRI feasibility outside academic infusion centers — early post-marketing series suggest higher symptomatic-ARIA rates in APOE ε4/ε4; systematic capture is pending.
4. SC starting dose in the real world. The July 2026 IQLIK 500 mg SC weekly starting therapy was approved on PK bridging plus IV outcomes; independent SC clinical-outcome data are not yet published.
5. Plasma assay governance and multi-product accuracy. With four FDA-cleared plasma products by August 2026, how to standardize reporting, prevent primary-care over-testing of cognitively unimpaired adults, and remediate the 2026 Lumipulse lot recall across previously tested cohorts remain unsettled.
6. Anticoagulation. AURs advise against; real-world AD patients frequently take chronic anticoagulation for AF. Whether short-term interruption or DOAC selection ever justifies treatment is an open question the AURs do not endorse.
7. Cost, coverage, and access equity. Medicare CED registry conditions, MRI capacity, and referral disparities remain rate-limiting.
8. Combination therapy (anti-amyloid + anti-tau; anti-amyloid + lifestyle) — trials ongoing; no practice-ready evidence.

Key clinical trials

What the general neurologist needs to know

Summary of evidence — not a protocol. No numeric dose, MRI interval, or titration schedule appears here; consult the current USPI before prescribing.

• Established: AD is a biological diagnosis under 2024 criteria; two anti-amyloid antibodies (lecanemab, donanemab) have traditional FDA approval; multiple plasma diagnostic aids (Lumipulse ratio, Elecsys pTau181 and pTau217 plasma, PrecivityAD2) are FDA-cleared for symptomatic adults in specialized care; ARIA is the class-defining adverse event and correlates with APOE ε4 dose.
• Uncertain: Long-term (>2 y) clinical benefit, treatment duration, patient-level clinical meaningfulness, and whether the July 2026 SC starting dose reproduces IV pivotal-trial benefit (approval rests on PK bridging and IV outcomes, not an independent SC clinical-outcome trial).
• Watch for: New headache, confusion, visual disturbance, gait instability, focal deficits, or seizure during the early months of treatment — obtain urgent MRI per institutional ARIA protocol and manage per the specific label. For lecanemab-treated patients presenting with acute ischemic stroke, remember the Leqembi USPI's explicit Warning against thrombolytic (tPA) use.
• Do not: Order plasma p-tau217, the pTau217/Aβ42 ratio, or plasma pTau181 in cognitively unimpaired adults; initiate an anti-amyloid antibody in a patient on chronic anticoagulation (both AURs advise against); assume plasma p-tau217 change confirms amyloid clearance after treatment (it does not, per the TB-ALZ 2 substudy); prescribe brexpiprazole for AD-associated agitation without disclosing its class Boxed Warning for increased mortality in elderly patients with dementia-related psychosis.

References

1. Jack CR Jr, Andrews JS, Beach TG, et al. Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup. Alzheimers Dement 2024. doi:10.1002/alz.13859
2. Palmqvist S, Tideman P, Mattsson-Carlgren N, et al. Plasma phospho-tau217 for Alzheimer's disease diagnosis in primary and secondary care using a fully automated platform. Nat Med 2025;31:2036–2043. doi:10.1038/s41591-025-03622-w
3. US Food and Drug Administration. Lumipulse G pTau217/β-Amyloid 1-42 Plasma Ratio 510(k) decision summary (K242706). May 2025.
4. US Food and Drug Administration. Elecsys pTau181 Plasma decision summary (K252163). October 2025.
5. US Food and Drug Administration. PrecivityAD2 510(k) decision (K253240). August 19, 2026.
6. US Food and Drug Administration. Elecsys pTau217 Plasma decision (K261686). August 2026.
7. US Food and Drug Administration. Class II Recall / Product Correction, Lumipulse G pTau217/β-Amyloid 1-42 Plasma Ratio (Res ID 217945). 2026.
8. US Food and Drug Administration. Lumipulse G β-Amyloid Ratio (1-42/1-40) CSF De Novo authorization (DEN200072).
9. US Food and Drug Administration. Elecsys β-Amyloid CSF authorization (K221842).
10. van Dyck CH, Swanson CJ, Aisen P, et al. Lecanemab in early Alzheimer's disease. NEJM 2023;388:9–21. doi:10.1056/NEJMoa2212948
11. Leqembi (lecanemab-irmb) US Prescribing Information (Eisai). Current label including 2026 update for IQLIK 500 mg SC starting therapy and IV/SC maintenance options. FDA reference label 761375s001.
12. US Food and Drug Administration. Approval of Leqembi IQLIK 500 mg SC weekly for starting therapy in Alzheimer's disease. July 2026.
13. Penner N, Rawal S, Aluri J, et al. Development of subcutaneous lecanemab: establishing comparability of subcutaneous and intravenous lecanemab formulations. Alzheimers Dement 2025;21(Suppl 5):e104694. doi:10.1002/alz.70859
14. Sims JR, Zimmer JA, Evans CD, et al. Donanemab in early symptomatic Alzheimer disease: TRAILBLAZER-ALZ 2. JAMA 2023;330:512–527. doi:10.1001/jama.2023.13239
15. Wang G, Cutter G, Oxtoby NP, et al. Modified titration of donanemab reduces ARIA-E risk: TRAILBLAZER-ALZ 6. Alzheimers Dement 2025;21:e70062. doi:10.1002/alz.70062
16. Kisunla (donanemab-azbt) US Prescribing Information (Lilly). Current label including 2025 titration update.
17. Cummings J, Apostolova L, Rabinovici GD, et al. Lecanemab: appropriate use recommendations. J Prev Alz Dis 2023. doi:10.14283/jpad.2023.30
18. Rabinovici GD, Selkoe DJ, Schindler SE, et al. Donanemab: appropriate use recommendations. J Prev Alz Dis 2025.
19. Cummings JL, Atri A, Feldman HH, et al. Efficacy and safety of oral semaglutide in early Alzheimer's disease: EVOKE and EVOKE+. Lancet 2026;407(10544):2167–2179. doi:10.1016/S0140-6736(26)00459-9
20. Lee D, Slomkowski M, Hefting N, et al. Brexpiprazole for the treatment of agitation in Alzheimer dementia: a randomized clinical trial. JAMA Neurol 2023;80:1307–1316. doi:10.1001/jamaneurol.2023.3810
21. Tariot PN, Cummings JL, Soto-Martin ME, et al. Trial of pimavanserin in dementia-related psychosis (HARMONY). NEJM 2021;385:309–319. doi:10.1056/NEJMoa2034634
22. Budd Haeberlein S, Aisen PS, Barkhof F, et al. Two randomized phase 3 studies of aducanumab in early Alzheimer's disease (EMERGE/ENGAGE). J Prev Alz Dis 2022;9:197–210. doi:10.14283/jpad.2022.30
23. Sperling RA, Donohue MC, Raman R, et al. Trial of solanezumab in preclinical Alzheimer's disease (A4). NEJM 2023;389:1096–1107. doi:10.1056/NEJMoa2305032
24. Rafii MS, Sperling RA, Donohue MC, et al. The AHEAD 3-45 study: design of a prevention trial for Alzheimer's disease. Alzheimers Dement 2023;19:1227–1233. doi:10.1002/alz.12748
25. Abushakra S, Hey JA, Aisen PS, et al. Valiltramiprosate in APOE ε4/ε4 early Alzheimer's disease: APOLLOE4. Drugs 2025;85:1455–1472. doi:10.1007/s40265-025-02250-5
26. Bateman RJ, Smith J, Donohue MC, et al. Long-term gantenerumab in dominantly inherited Alzheimer's disease: DIAN-TU open-label extension. Lancet Neurol 2025;24:316–330.
27. Collins EC, Zimmer JA, Sims JR, et al. Plasma p-tau217 for detecting treatment-related amyloid clearance after donanemab: TRAILBLAZER-ALZ-2 substudy. Alzheimers Dement 2026;22:e71740. doi:10.1002/alz.71740
28. Ferreira PCL, Karikari TK, Bellaver B, et al. Evaluating plasma p-tau217 as a clinical trial endpoint in Alzheimer's disease. Neurology 2026;106:e214441.
29. Barthélemy NR, Bateman RJ, et al. Predicting onset of symptomatic Alzheimer's disease with plasma p-tau217 clocks. Nat Med 2026.
30. Alzheimer's Association. 2026 Alzheimer's Disease Facts and Figures. Alzheimers Dement 2026.
31. Lynch S, et al. ABBV-916 phase 1b/2 in early Alzheimer's disease. Alzheimers Dement 2026;22:e71715. doi:10.1002/alz.71715
32. Pascual-Lucas M, Lacosta AM, et al. ABvac40 long-term safety and immunological memory (AB1601). Alzheimers Dement 2026;22:e71746. doi:10.1002/alz.71746
33. Sperling RA, Jack CR Jr, Aisen PS, et al. ARIA update. Lancet Neurol 2024.
34. Graff-Radford J, Yong KXX, Apostolova LG, et al. New insights into atypical Alzheimer's disease. Lancet Neurol 2021;20:222–234.
35. Morris MC, Tangney CC, Wang Y, et al. MIND diet slows cognitive decline with aging. Alzheimers Dement 2015;11:1015–1022. doi:10.1016/j.jalz.2015.04.011
36. Barnes LL, Dhana K, Liu X, et al. Trial of the MIND diet for prevention of cognitive decline in older persons. NEJM 2023;389:602–611. doi:10.1056/NEJMoa2302368
37. Salloway S, Sperling R, Fox NC, et al. Two phase 3 trials of bapineuzumab in mild-to-moderate Alzheimer's disease. NEJM 2014;370:322–333.