Reviews/Alzheimers Disease Treatment/Anti-amyloid antibodies cause amyloid-related imaging abnormalities
Amyloid Immunotherapy · AD-AMY-03 · v2
Several large, blinded trials in which patients were randomly assigned to an amyloid-targeting antibody or a dummy infusion found these brain-imaging changes far more often in the treated patients than in the untreated ones. This pattern held across four different antibodies, so it is not a quirk of one drug. Because patients were assigned by chance, the much higher rate in treated patients points to the drug itself as the cause rather than to some pre-existing difference between the groups. A real-world patient-record study and a laboratory study in mice pointed the same way. These imaging changes also arise on their own in untreated Alzheimer patients, but at a much lower rate, so the antibodies add substantially to that background. The panel treated this as a clear drug-caused adverse effect.
Four placebo-controlled randomized trials of distinct antibodies report the imaging abnormalities far above placebo: aducanumab 41.3% ARIA versus 2.7% placebo ARIA-E [1], donanemab ARIA-E 24.4% versus 1.9% [19], bapineuzumab ARIA-E 15.8% versus 0.8% [22], and lecanemab ARIA-E 6.2% versus 1.4% [18]. A phase 2 lecanemab trial found ARIA-E incidence below 10% and tied it to drug exposure [17]. A real-world comparison of monoclonal versus non-monoclonal treatment found a cumulative hazard ratio of 4.65 (95% CI 3.77-5.73) [13]. A mouse study showed that an antibody binding to cerebral amyloid angiopathy drove complement activation and microhemorrhages [29]. The randomized design is what lets the large treated-versus-placebo gap be read as caused by the antibody rather than merely associated. The contested point: The imaging abnormalities are not exclusive to treatment: they occur spontaneously in untreated Alzheimer disease at placebo rates of 1.4% to 2.7% [1][18][19], so the antibody is an added cause rather than the sole one. The underlying biological mechanism is shown directly only in mice [29] and in-vitro or ex-vivo binding work [3][11][28], not confirmed in humans.
All 36 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Alzheimers Disease Treatment, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
The claim verb is 'cause,' yet ARIA also arises spontaneously in untreated AD (placebo ARIA-E 1.4-2.7% in [1][18][19]), so antibodies are not the sole cause. However randomization in the placebo-controlled trials attributes the large excess incidence causally to the antibody, so the claim as worded holds.
Several included sources are indirect or surrogate-level: [29][6] are mouse models, [28][3][11] are in-vitro/ex-vivo binding studies, and [5] is an active vaccine (endogenous antibody, not a therapeutic monoclonal). These must not carry the causal verdict, which rests on the human RCTs.
Where it remains uncertain. The abnormalities also arise spontaneously in untreated patients [1][18][19], and the causal mechanism is demonstrated directly only in a mouse model [29] and lab binding studies [3][11], not in humans.
Aleatory
Incidence tracks individual factors such as APOE ε4 status and baseline microhemorrhages [19], so occurrence in any one patient is probabilistic rather than fixed.
Epistemic
The complement and cerebral amyloid angiopathy pathway is demonstrated directly only in mice [29] and lab binding studies [3][11][28]; why lecanemab causes fewer events than other antibodies remains unknown [11].
Inferential
The hazard ratio of 4.65 [13] comes from a propensity-matched electronic-record cohort, not randomization, so residual confounding cannot be excluded.
The load-bearing papers from a corpus of 36, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Aducanumab phase 3 RCTs; 41.3% ARIA vs 2.7% placebo ARIA-E establishes treatment-driven ARIA.
“During the placebo-controlled period, 425 of 1029 patients (41.3%) experienced ARIA, with serious cases occurring in 14 patients (1.4%).”
Panel auditInternal validity: high · scope of inference: Patients with MCI or mild AD dementia receiving aducanumab.Key limitations: trials terminated early for futility · high withdrawal · reports ARIA incidence not formal statistical test
Real-world comparative evidence: monoclonal (amyloid antibody) treatment carries ~4.6x ARIA hazard vs non-monoclonal.
“The monoclonal group had significantly higher cumulative hazard of ARIA (hazard ratio = 4.65, 95% CI = 3.77-5.73; P < 0.001) compared with the nonmonoclonal group.”
Panel auditInternal validity: medium · scope of inference: AD patients on monoclonal vs nonmonoclonal drugs in an EHR dataset.Key limitations: retrospective observational · EHR coding reliance · residual confounding
Lecanemab phase 2 RCT; ARIA-E incidence modelled against drug exposure, supporting dose-dependent causation.
“In the phase 2 core and OLE, there was a low incidence of ARIA-E (<10%), with <3% symptomatic cases.”
Panel auditInternal validity: medium · scope of inference: Early AD patients receiving lecanemab.Key limitations: no placebo ARIA rate stated · no p-value · low overall ARIA-E incidence
Measures antibody binding patterns to plaque and vasculature, not any ARIA outcome, so direction on the claim is not determinable.
Panel auditInternal validity: not assessable from abstract · scope of inference: In-vitro binding behaviour of two anti-Aβ antibodies in mouse brain.Key limitations: ex vivo mouse tissue · binding surrogate not ARIA · fixation artefacts noted
4 of 36 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Multiple independent randomized placebo-controlled trials across different antibodies show a large, consistent excess of ARIA attributable to treatment, making this among the most firmly established drug-adverse-event causal relationships in the corpus.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Human mechanism confirmation. The complement and CAA mechanism that explains why the antibodies cause the abnormalities is demonstrated only in mice [29] and lab binding studies [3][11], leaving the human causal pathway inferred rather than observed. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: Human neuropathological or biomarker studies linking antibody binding to cerebral amyloid angiopathy, complement activation and incident imaging abnormalities.
Open question: Does the antibody-binding-to-CAA and complement mechanism demonstrated in mice operate in humans? Why it is still open: Mechanistic evidence is confined to a mouse model [29] and lab binding studies [3][11][28]; no human tissue study in the corpus confirms it.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19