Anti-dementia drugs in 2026: what works

A June 2026 review in Deutsche Medizinische Wochenschrift by Marlena Schnieder and Christine von Arnim at the University of Gottingen surveys the current state of drug treatment for dementia in older adults^[1]. The review covers three categories: the long-established cholinesterase inhibitors and memantine, the newer amyloid-targeting antibodies lecanemab and donanemab, and the GLP-1 receptor agonist semaglutide, whose two Phase 3 Alzheimer's trials reported negative results earlier this year.

For PeptideMethods readers, the review matters because it puts the semaglutide failure in context. The excitement about GLP-1 drugs in neurodegeneration ran high through 2024 and into 2025. The EVOKE and EVOKE+ trial results closed that chapter for now.

The old guard: cholinesterase inhibitors and memantine

Donepezil, rivastigmine, and galantamine remain the standard first-line medications for Alzheimer's disease. Schnieder and von Arnim describe their cognitive benefits as "modest" but note that the side-effect profile is manageable for most patients^[1]. These drugs do not alter the course of the disease. They improve symptoms for a period, and the effect fades as the underlying neurodegeneration progresses.

Memantine, an NMDA receptor antagonist, is typically added in moderate-to-severe stages. Off-label use of these drugs in vascular dementia is common, though the evidence base is thinner^[1].

Amyloid-targeting antibodies: lecanemab and donanemab

The review covers the two amyloid-clearing monoclonal antibodies that have reached regulatory approval. Lecanemab (Leqembi, Eisai/BioArb) received full FDA approval in July 2023 following the CLARITY AD trial, which showed a 27% slowing of cognitive decline on the CDR-SB scale over 18 months in 1,795 participants with early Alzheimer's^[2]. The EMA approved lecanemab for the EU after a protracted review that included a formal appeal.

Donanemab (Kisunla, Eli Lilly) demonstrated a 35% slowing of decline on the iADRS in the TRAILBLAZER-ALZ 2 trial of 1,736 participants^[3]. The FDA approved it in July 2024. The EMA initially rejected donanemab in March 2025 over safety concerns, particularly the 36.8% rate of amyloid-related imaging abnormalities (ARIA), but reversed course in July 2025, recommending approval for patients who are ApoE4 non-carriers or heterozygotes.

Both antibodies require amyloid-PET or CSF confirmation before treatment, regular MRI monitoring for ARIA, and infusion infrastructure. The practical access barriers are real, especially outside specialised memory clinics. The review notes that thorough geriatric assessment is important for balancing the risks of these treatments against the patient's overall health status.

Semaglutide: the GLP-1 bet that did not pay off

The most notable finding for this site's readers: Schnieder and von Arnim note that "the failure of GLP-1 analogs (e.g., semaglutide) in recent trials highlights the challenges in repurposing metabolic agents for neurodegeneration"^[1].

They are referring to EVOKE and EVOKE+, two parallel Phase 3 trials that enrolled 3,808 participants aged 55 to 85 with mild cognitive impairment or mild dementia due to Alzheimer's disease. Oral semaglutide 14 mg showed no difference from placebo on the primary endpoint (change in CDR-SB at week 104) in either trial^[4]. A pooled analysis also failed to show delayed progression from MCI to dementia over 156 weeks.

The disconnect between mechanism and outcome was striking. In a CSF substudy of roughly 200 patients, semaglutide reduced core Alzheimer's biomarkers by up to 10% and cut C-reactive protein by about 30%^[4]. None of that translated into clinical benefit. Novo Nordisk discontinued the extension periods of both trials after reviewing these results.

This is a meaningful negative result. Observational data and preclinical work had suggested that GLP-1 agonists might protect against neurodegeneration through anti-inflammatory and vascular pathways. The EVOKE programme was the first rigorous test of that hypothesis, and it did not hold up.

Non-pharmacological approaches

The review also covers non-drug interventions. It highlights 150 minutes per week of moderate exercise as an evidence-supported recommendation, along with cognitive-training digital applications and multimodal intervention programmes such as FINGER and POINTER^[1]. The emphasis on early diagnosis and integrated geriatric assessment runs through the paper.

What this means for readers following GLP-1 research

The semaglutide-for-Alzheimer's story is instructive. Observational signals can look promising. Mechanistic arguments can sound plausible. But large, well-designed trials remain the test that matters, and EVOKE and EVOKE+ did not pass it. Readers who track semaglutide for its metabolic indications should be aware that the neurodegeneration story has, for now, been closed by the data.

For amyloid-targeting therapies, the picture is different. Lecanemab and donanemab represent genuine, if modest, advances: the first treatments shown to slow the biological progression of Alzheimer's in randomised trials. Access remains the bottleneck, and the safety monitoring requirements are non-trivial. But the direction of travel is clear.

This article is for informational purposes only and does not constitute medical advice. If you or someone you know has concerns about cognitive decline, consult a qualified healthcare provider.