Reviews/Alzheimers Disease Treatment/Memantine helps moderate-to-severe Alzheimer's disease
Symptomatic · AD-SYM-02 · v2
Four separate high-quality drug trials in people with moderate to severe Alzheimer's, each comparing memantine against a dummy pill, found that people on memantine scored better on tests of thinking and memory than those on placebo [21][24][36][1]. The result showed up repeatedly, whether memantine was given on its own [21] or added to another dementia drug [24][36]. But the size of the benefit was small, and in one trial it fell short of the trial's own bar for a difference patients would actually notice [1]. In another trial, the memantine group barely changed while the placebo group got worse, so the drug looks more like it slows decline than truly improves thinking [36]. The panel judged the claim well supported but held it back from the top tier because the effect is modest and sits near the edge of clinical meaningfulness.
Four independent double-blind placebo-controlled RCTs in the exact population show a statistically significant memantine advantage on a cognitive instrument: the Severe Impairment Battery in [21], [24] (difference 2.6, 95% CI 1.0 to 4.2, p=0.001) and [36] (memantine +0.9 vs placebo -2.5, P<.001), and the SMMSE in [1] (+1.2 points, 95% CI 0.6 to 1.8, P<0.001). The signal holds across memantine monotherapy [21] and cholinesterase-inhibitor add-on designs [24][36]. Magnitudes are small on wide-range scales, and in [1] the +1.2-point SMMSE gain sits below the trial's own 1.4-point minimum clinically important difference. In [36] the memantine group held roughly stable while placebo declined, indicating attenuated deterioration rather than frank gain. The contested point: The division is over whether a replicated, statistically significant signal of small magnitude amounts to the claim's word 'improves'. In [36] memantine barely moved (+0.9) while placebo fell (-2.5), and in [1] the gain fell below the pre-specified MCID, so several panel members read the effect as slowed decline near the threshold of clinical meaning rather than genuine cognitive improvement.
All 38 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.
In DOMINO [1] the memantine SMMSE benefit (1.2 points) falls below the trial's own pre-specified minimum clinically important difference of 1.4 points, so a significant p-value attaches to a sub-threshold effect.
'Improves cognition' overstates the data: in [36] memantine barely changed (+0.9) while placebo declined (-2.5), so the measured effect is largely slowed deterioration rather than genuine improvement.
Where it remains uncertain. The magnitude is small and near-threshold: [1] shows +1.2 SMMSE points below its own 1.4-point MCID, and [36] shows memantine holding stable (+0.9) while placebo declined (-2.5), so the effect is partly slowed decline rather than improvement.
Aleatory
SIB differences of roughly 2.6-3.4 points on a 0-100 scale and a 1.2-point SMMSE gain are small relative to measurement variability [24][1].
Epistemic
The study-type filter removed the Cochrane review [35] and the Schneider meta-analysis [38], so the authoritative meta-analytic view is absent from eligible evidence.
Inferential
In [36] memantine held stable (+0.9) while placebo declined (-2.5), so whether 'improves' is warranted turns on reading attenuated decline as improvement.
The load-bearing papers from a corpus of 38, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Independent pragmatic RCT (N=295) confirms a significant SMMSE cognitive benefit, but the 1.2-point effect sits below the trial's 1.4-point minimum clinically important difference.
Panel auditInternal validity: high · scope of inference: Community-dwelling patients with moderate-to-severe AD already on donepezil, over 52 weeks.Key limitations: 1.2-point gain is below the 1.4-point SMMSE minimum clinically important difference · all patients had prior donepezil exposure
Memantine monotherapy significantly outperformed placebo on the SIB cognitive measure in moderate-to-severe AD (N=252).
“the Severe Impairment Battery (P<0.001 with the last observation carried forward, P=0.002 for observed cases)”
Panel auditInternal validity: high · scope of inference: Moderate-to-severe AD patients over 28 weeks, US centres.Key limitations: SIB was a secondary endpoint (primaries were global and function) · high early discontinuation
Extended-release memantine on a cholinesterase inhibitor gave a significant SIB cognitive benefit in moderate-to-severe AD (N=677).
“memantine-treated patients significantly outperformed placebo-treated patients on the SIB (least squares mean difference [95 % CI] 2.6 [1.0, 4.2]; p = 0.001)”
Panel auditInternal validity: high · scope of inference: Moderate-to-severe AD on cholinesterase inhibitors over 24 weeks.Key limitations: functional ADCS-ADL19 co-endpoint not significant · all on background cholinesterase inhibitor
3 of 38 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.
Direct, replicated RCT evidence in the exact population supports a modest cognitive benefit, but small magnitude and near-threshold clinical significance hold it short of Established.
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.
Effect magnitude against a clinical-importance threshold. Whether the cognitive benefit is patient-meaningful, not just significant, is unresolved; [1] shows a sub-MCID gain. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: An RCT powered to detect a difference exceeding a pre-specified cognitive MCID as its primary endpoint.
Open question: Does memantine produce a cognitive benefit that exceeds a pre-specified minimum clinically important difference in moderate-to-severe AD? Why it is still open: [1] shows a significant +1.2-point SMMSE gain that falls below its own 1.4-point MCID.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19