Reviews/Anti Amyloid Antibodies/Anti-amyloid effects exceed minimal clinically important difference
Cme · AMAB-CME-01 · v2
Borderline: within 4 points of the Doubtful/Refuted boundary, so a re-grade could land either side.
This claim asks whether the slowing of decline produced by anti-amyloid antibodies is large enough to clear the bar that patients and clinicians would call a meaningful difference. The corpus supplies both halves of that comparison: the antibody trials report slowing of about a third to just under half a point on the CDR-SB scale [1][2][3], while the corpus's own benchmarks for a meaningful change sit several times higher [15][13]. On a straight reading, the observed effects are well short of those benchmarks, and two of the three antibodies did not even produce a clear benefit [2][3]. The panel noted that meaningfulness is defended in the corpus by other routes, such as extra months of independent living [14] and preserved quality of life [6], but none of those directly show the effect exceeding a stated threshold. There is also a genuine dispute about whether a group-average difference should be compared against a within-patient threshold at all. Weighing the direct numbers against that framing dispute, the panel did not find the claim supported.
The three phase 3 programmes report CDR-SB between-group differences of -0.45 for lecanemab [1], -0.39 in EMERGE versus a null 0.03 in ENGAGE for aducanumab [2], and non-significant -0.31 and -0.19 across the two GRADUATE gantenerumab trials [3]. Against these, the corpus's meaningful-change benchmarks are far larger: a within-patient CDR-SB range of roughly 1.5 to 2.5 points in early AD [15] and test-specific MCIDs anchored to CDR-SB changes of 0.5 to 1.0 [13]. The largest observed effect is several-fold below the within-patient benchmark. Meaningfulness is argued through other framings, including about 10 months of extended IADL independence with lecanemab [14] and 49 to 56 percent less decline on quality-of-life measures [6], but these do not test whether the effect exceeds an MCID. The contested point: The division is whether a group-level mean difference of about 0.3 to 0.45 on CDR-SB can legitimately be set against within-patient meaningful-change thresholds of 1.5 to 2.5 [15], which are a different quantity. Cummings [10] frames disease-modifying effects as progressive, so an MCID not reached early may be reached and later exceeded with continued therapy, while the plain numerical comparison in the corpus shows every observed effect below every threshold.
All 17 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 Anti Amyloid Antibodies, 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 compares a group-level mean difference (~0.45) to within-patient MCIDs (1.5-2.5, [15]); these are different quantities and the corpus contains no analysis showing an antibody effect reaching an MCID on a like-for-like basis.
Every observed effect is below every MCID threshold in the corpus: -0.45 [1], -0.39 [2], -0.31 [3] versus 1.5-2.5 [15] and ~1-2 [13]. Nothing exceeds the MCID; on the plain numerical reading the claim is contradicted.
Where it remains uncertain. The direct comparison is a group mean against a within-patient threshold [15], and Cummings argues disease-modifying effects accrue and may later exceed an MCID [10], which the single-timepoint trial data cannot settle.
Aleatory
The same molecule and class gave opposing results across paired trials, -0.39 in EMERGE versus null 0.03 in ENGAGE [2] and two non-significant GRADUATE readouts [3], so sampling variability alone leaves the effect magnitude unstable.
Epistemic
The corpus holds effect sizes and MCID benchmarks but no single analysis that compares an antibody effect against an MCID on the same quantity [1][15].
Inferential
The verdict's central comparison places a between-group mean difference [1] against within-patient meaningful-change thresholds [15], two different quantities the corpus never reconciles.
The load-bearing papers from a corpus of 17, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Largest antibody effect on the primary outcome, -0.45.
“The adjusted least-squares mean change from baseline at 18 months was 1.21 with lecanemab and 1.66 with placebo (difference, -0.45; 95% confidence interval [CI], -0.67 to -0.23; P<0.001).”
Panel auditInternal validity: high · scope of inference: Adults 50-90 with early Alzheimer's disease (MCI or mild dementia) with confirmed amyloid, over 18 months.Key limitations: no MCID comparison in abstract · ARIA and infusion reactions noted · 18-month horizon only
Second molecule: small positive effect in one trial, null in its twin.
“The primary endpoint was met in EMERGE (difference of -0.39 for high-dose aducanumab vs placebo [95% CI, -0.69 to -0.09; P=.012; 22% decrease]) but not in ENGAGE (difference of 0.03, [95% CI, -0.26 to 0.33; P=.833; 2% increase]).”
Panel auditInternal validity: medium · scope of inference: Patients 50-85 with MCI due to AD or mild AD dementia with confirmed amyloid, over 76 weeks.Key limitations: trials halted for futility · discordant results between two trials · no MCID comparison
Independent cohort-derived MCID benchmarks, all above observed effects.
“We identified potential triangulated MCIDs (cognitively unimpaired; MCI) on a range of cognitive test outcomes: MMSE -1.5, -1.7; ADAS delayed recall 1.4, 1.1”
Panel auditInternal validity: medium · scope of inference: MCID thresholds for cognitive tests in cognitively unimpaired and MCI populations.Key limitations: no treatment data · observational cohort · MCIDs are context-specific
Third molecule: non-significant, below any MCID.
“The change from baseline in the CDR-SB score at week 116 was 3.35 with gantenerumab and 3.65 with placebo in the GRADUATE I trial (difference, -0.31; 95% confidence interval [CI], -0.66 to 0.05; P = 0.10)”
Panel auditInternal validity: high · scope of inference: Adults 50-90 with MCI or mild dementia due to AD with amyloid evidence, over 116 weeks.Key limitations: non-significant primary endpoints · no clinical decline slowing · symptomatic ARIA-E 5.0%
4 of 17 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.
The direct comparison the corpus affords places antibody effects several-fold below cited MCID thresholds, with residual confidence only because the group-versus-within-patient framing is contested and accrual/meaningfulness signals exist.
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.
No within-patient responder analysis against an MCID. The verdict turns on comparing a group mean [1] with a within-patient threshold [15], which are different quantities; the corpus offers no responder analysis that puts them on the same footing. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: A within-patient responder analysis of the antibody trials applying the CDR-SB meaningful-change thresholds [15] to individual-level change to report the share of treated patients exceeding the MCID.
Open question: What proportion of antibody-treated patients achieve a within-patient CDR-SB change exceeding the meaningful-change threshold? Why it is still open: The corpus reports only group-mean differences [1][2][3] and separately a within-patient threshold [15], never a responder-level comparison.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19