As of 2026, two approved passive immunotherapies — lecanemab and donanemab — demonstrate statistically significant but modest slowing of cognitive decline in early Alzheimer's disease: 27% and 35% respectively on primary clinical scales versus placebo. Effect sizes are real but incremental, confined to amyloid-confirmed early-stage disease, and accompanied by clinically significant ARIA rates that constrain patient selection.
How Do Amyloid-β–Targeting Peptide Immunotherapies Work at the Molecular Level?
Passive immunotherapies deliver exogenous monoclonal antibodies that bind distinct Aβ epitopes. Lecanemab targets soluble protofibrils and aggregated Aβ species, while donanemab selectively recognises N-terminally truncated pyroglutamate-modified Aβ (AβpE3), a post-translationally modified form enriched in mature plaques. Antibody–antigen complexes are cleared via Fc-receptor–mediated microglial phagocytosis and perivascular drainage.
The epitope distinction carries mechanistic weight. Lecanemab's preferential binding to soluble protofibrils — the species most strongly implicated in synaptic toxicity — is hypothesised to interrupt neurotoxic signalling upstream of plaque consolidation. Donanemab's pyroglutamate-Aβ selectivity concentrates activity at established fibrillar deposits, producing the most rapid amyloid PET reductions observed in any Phase 3 trial to date.
Active peptide immunotherapy approaches, including UB-311 (a synthetic Aβ1–14 B-cell epitope vaccine), aim to elicit endogenous polyclonal antibody responses without the T-cell–mediated meningoencephalitis that terminated the AN-1792 programme in 2002. UB-311's Phase 2a data confirmed 96–100% immunogenicity with a favourable safety profile, though the trial was not powered to detect cognitive efficacy signals.
What Did the CLARITY AD Trial Establish About Lecanemab's Cognitive Effect Size?
CLARITY AD enrolled 1795 participants with early Alzheimer's disease over 18 months and demonstrated that lecanemab 10 mg/kg biweekly reduced CDR-Sum of Boxes decline by 27% versus placebo, with an absolute between-group difference of nearly half a point (statistically significant at p<0·001). Amyloid PET burden fell by a mean 59 Centiloids in the treatment arm.
The absolute CDR-SB difference at 18 months sits below the 1–2 point threshold often cited as clinically meaningful by dementia specialists. This generates ongoing debate about whether statistical significance translates to functional benefit perceptible to patients and caregivers. Secondary endpoints including ADAS-Cog14, ADCS-MCI-ADL, and MMSE all trended in the same direction, providing convergent validity.
ARIA-E (edema/effusion) occurred in approximately 13% of lecanemab-treated participants versus under 2% on placebo, while ARIA-H (microhemorrhages/hemosiderosis) occurred in approximately 17% versus 9%. Most ARIA events were radiographically detected and asymptomatic, but serious symptomatic cases did occur. APOE ε4 homozygotes experienced ARIA at roughly double the rate of non-carriers, a pharmacogenomic stratification now central to prescribing guidance.
What Did TRAILBLAZER-ALZ 2 Reveal About Donanemab's Efficacy and Tau-Stratified Response?
TRAILBLAZER-ALZ 2 enrolled 1736 participants over 76 weeks and showed donanemab slowed iADRS decline by approximately 35% in the low/medium tau subgroup (p<0·001), with CDR-SB slowing reaching 36% in that group. Amyloid plaque reduction reached 87 Centiloids by week 76, and 52% of participants achieved sub-threshold amyloid clearance enabling protocol-specified treatment discontinuation.
The tau-stratification design was a deliberate methodological choice: participants were enrolled with confirmed amyloid positivity and stratified by flortaucipir PET into low/medium versus high tau burden cohorts. The high-tau subgroup showed substantially attenuated benefit, consistent with the hypothesis that downstream tau pathology limits the ceiling of amyloid-clearance benefit.
ARIA-E occurred in approximately 24% of donanemab-treated participants versus 2% on placebo, and ARIA-H in approximately 31% versus 14% (Schreiber and colleagues, 2025, PMC12660453). Three deaths in the donanemab arm were adjudicated as ARIA-related, establishing a rare but non-negligible fatal risk. Indirect comparison analyses suggest donanemab carries a higher ARIA burden than lecanemab, particularly in APOE ε4 carriers.
How Do Plasma and CSF Biomarkers Track Treatment Response in These Trials?
Both trials documented significant reductions in plasma phospho-tau 181 (p-tau181) and p-tau217, alongside amyloid PET clearance, establishing that peripheral biomarker changes co-occur with central amyloid reduction. Plasma p-tau217 has emerged as the most sensitive single marker for tracking Aβ immunotherapy response, with changes detectable before cognitive endpoint divergence.
Cerebrospinal fluid Aβ42/40 ratios and total tau levels shifted in the expected direction in both programmes, but the magnitude of CSF change did not correlate linearly with cognitive outcome. This underscores that amyloid clearance is necessary but not sufficient for functional preservation. Neurodegeneration markers including neurofilament light chain (NfL) showed attenuated rise in treatment arms, suggesting downstream neuroprotective effects beyond direct plaque removal.
The field is moving toward plasma p-tau217 as a screening and monitoring tool that could replace PET confirmation in some settings, reducing cost and access barriers. A 2025 PMC review by Sha and colleagues (PMC12220716) identifies plasma biomarker normalisation as a secondary endpoint in multiple ongoing Phase 3 extensions, with regulatory agencies increasingly receptive to biomarker-supported labelling claims.
Where Does Active Peptide Vaccine Immunotherapy Stand in 2026?
Active Aβ peptide vaccines — UB-311, ABvac40, and the discontinued CAD106 — aim to generate endogenous antibody responses against defined epitopes, avoiding the infusion burden of passive therapy. As of 2026, no active vaccine has completed a Phase 3 efficacy trial, and UB-311 leads with 96% immunogenicity in Phase 2a but underpowered cognitive signal data.
UB-311 (Vaxxinity) uses a synthetic Aβ1–14 peptide conjugated to a UBITh® T-helper epitope, eliciting IgG responses without activating autoreactive T cells — the mechanism implicated in AN-1792's meningoencephalitis. Phase 2a results confirmed safety and durable antibody titres in mild AD patients, with secondary cognitive endpoints trending favourably but not reaching statistical significance in the small cohort.
ABvac40 targets the C-terminus of Aβ40, a less-studied epitope region. Phase 2 data presented in 2024 showed significant cognitive stabilisation in participants who mounted high antibody titres, though the trial was not powered for a primary cognitive endpoint. The divergence between high- and low-responder subgroups suggests antibody titre thresholds may be a critical determinant of active vaccine efficacy.
Is the Observed Cognitive Slowing Clinically Meaningful or Statistically Nominal?
The clinical meaningfulness debate centres on whether a sub-half-point CDR-SB difference (lecanemab) or a 35% slowing rate (donanemab) translates to perceptible functional benefit. Regulatory agencies accepted these effect sizes for approval, though independent commentators note that neither trial demonstrated a patient-reported outcome improvement and that absolute between-group differences remain small relative to natural disease variability.
A 2025 update by Wu and colleagues (PMC12637128) frames the approvals as a "change in thinking" — the first demonstration that targeting a defined pathological substrate can alter the trajectory of AD, even if the magnitude is modest. The argument is that early-stage intervention, before tau spread and neurodegeneration become self-sustaining, is the mechanistically correct window, and that incremental slowing compounds over years of treatment.
Critics point to the absence of a validated minimal clinically important difference (MCID) for CDR-SB in early AD. The exclusion of APOE ε4 homozygotes from some real-world prescribing contexts due to ARIA risk and the high cost of biweekly infusion regimens further complicate the benefit-risk calculus outside controlled trial conditions. The field awaits longer-term extension data to determine whether early slowing translates to delayed milestone events such as loss of independence.
Which Patient Characteristics Predict Differential Benefit or Risk in 2026?
Confirmed amyloid positivity (PET or CSF), early disease stage (MCI or mild dementia), low-to-medium tau burden, and APOE ε4 non-carrier status collectively define the subgroup with the most favourable benefit-risk profile. APOE ε4 homozygotes face disproportionate ARIA risk and require enhanced monitoring or exclusion in most current prescribing frameworks.
Tau burden stratification, validated in TRAILBLAZER-ALZ 2, is emerging as a pre-treatment selection criterion: participants with high tau pathology derived substantially less cognitive benefit despite equivalent amyloid clearance. This finding supports the hypothesis that tau-mediated neurodegeneration, once established, proceeds independently of amyloid burden and limits the therapeutic ceiling of Aβ-targeted approaches.
Age interacts with both efficacy and safety: TRAILBLAZER-ALZ 2 subgroup analyses showed participants under 75 achieved 48% iADRS slowing versus attenuated effects in those aged 75 or older. Renal function, anticoagulant use, and prior microhemorrhage burden are additional ARIA risk modifiers addressed in clinical practice guidelines published by the Alzheimer's Association in 2023 and updated in 2024. How Much Does Semaglutide 2.4 mg Reduce Major Cardiovascular Events in Non-Diabetic Patients With Established CVD in 2026? Is PT-141 Safe for Patients With Cardiovascular Comorbidities in 2026? What Are the Evidence-Based Dosing Protocols for Retatrutide in the TRIUMPH Phase 3 Trial Versus Tirzepatide in 2026?