An Update You Deserved: The EVOKE Trials, What the Data Shows, and Why Hope Remains

Several months ago, I published a post on this blog about the emerging research into GLP-1 receptor agonists — drugs like semaglutide (Ozempic, Wegovy) — and their potential role in Alzheimer's disease prevention and treatment. I was careful then to frame the science accurately: the observational data was compelling, the biological rationale was sound, and the Phase 3 EVOKE and EVOKE+ trials were the rigorous test the field needed. I promised to update when the results arrived.

They have arrived. And I owe an honest accounting of what they show — the disappointing, the nuanced, and the genuinely encouraging.

What the EVOKE Trials Were

To briefly recap: EVOKE and EVOKE+ were two of the largest Alzheimer's clinical trials ever conducted — randomized, double-blind, placebo-controlled Phase 3 trials enrolling a combined 3,808 adults aged 55 to 85 across 40 countries. All participants had biomarker-confirmed early-stage Alzheimer's disease, either at the MCI or mild dementia stage. They were assigned to receive once-daily oral semaglutide at 14 mg or placebo for a primary treatment phase of 104 weeks, with a planned one-year extension.

The primary endpoint was change in CDR-SB — the Clinical Dementia Rating Sum of Boxes — the most rigorous and widely accepted yardstick for measuring cognitive and functional decline in Alzheimer's trials. This was, by every methodological measure, the gold standard test the field had been waiting for.

The Primary Results: What Did Not Happen

On November 24, 2025, Novo Nordisk announced the topline results. They were negative.

Oral semaglutide did not demonstrate superiority over placebo on the CDR-SB primary endpoint in either trial. The treatment and placebo curves, when presented at the Clinical Trials on Alzheimer's Disease (CTAD) conference in December, were, in the words of one researcher who attended, essentially indistinguishable over the full three years of the study.

The secondary endpoints — including the Montreal Cognitive Assessment (MoCA), the ADAS-Cog13, the MMSE, and functional measures of daily living — told the same story. No meaningful difference between the groups. The pooled analysis across both trials found that semaglutide did not delay progression from MCI to mild dementia, with a hazard ratio of 0.96 — essentially a coin flip. Based on these results, Novo Nordisk discontinued the one-year extension period and formally ended the program.

Full results were subsequently published in The Lancet and presented at the AD/PD 2026 International Conference on Alzheimer's and Parkinson's Diseases in March 2026 in Copenhagen, where the complete data package was laid out for the scientific community for the first time.

I want to be direct about what this means for my patients: semaglutide is not a treatment for Alzheimer's disease. It should not be sought out, requested, or used off-label for that purpose. It remains an excellent, well-validated medication for Type 2 diabetes and obesity — and its role in those conditions is undiminished by these results.

What Did Happen: The Biomarker Story

Here is where the data becomes considerably more interesting — and where I think the interpretation requires more than a simple "it failed" headline.

Embedded within the EVOKE trials was a rigorous CSF sub-study involving 199 patients who underwent cerebrospinal fluid sampling. In this sub-study, semaglutide produced nominally significant reductions of approximately 10% in several key Alzheimer's biomarkers compared to placebo. Specifically, the drug reduced CSF levels of p-tau-181, p-tau-217, non-phosphorylated-ta-181, and non-phosphorylated-tau-205 — all markers directly tied to the neurofibrillary tangle pathology that drives neurodegeneration in Alzheimer's. It also reduced YKL-40, a marker of neuroinflammation, as well as total tau and neurogranin, markers of neurodegeneration and synaptic injury. Separately, plasma hs-CRP — a systemic inflammatory marker — was meaningfully reduced across the full trial population.

Let me translate that into plain language. The drug was doing something biological. It was measurably reducing the toxic proteins and inflammatory signals that define Alzheimer's pathology. It was touching the right targets.

But that 10% reduction, while statistically real, was not large enough to produce a clinically detectable benefit in patients who already had established symptomatic disease. The biology was moving — just not enough, and perhaps not at the right stage.

Why the Gap Between Biology and Clinical Benefit?

This is the question the field is now grappling with seriously, and I think the most thoughtful interpretation comes from a critical distinction that experts highlighted at both CTAD and AD/PD.

The observational studies that originally generated excitement about GLP-1 agonists — including a landmark analysis of over one million patients with Type 2 diabetes showing a 40% to 70% reduction in first-time Alzheimer's diagnoses with semaglutide — were studying a fundamentally different question. They were asking: does metabolic management with a GLP-1 agonist reduce the risk of developing Alzheimer's in people with diabetes and obesity? That is a prevention question.

The EVOKE trials asked a different question entirely: can semaglutide slow the progression of disease in people who already have confirmed Alzheimer's pathology? That is a treatment question. And the biology of the disease may make those two questions very different to answer. By the time a patient has symptomatic, amyloid-confirmed Alzheimer's, the cascade of neurodegeneration is already deeply entrenched. A 10% nudge in biomarkers may not be sufficient to meaningfully redirect a process that has been building for fifteen to twenty years.

As one leading researcher put it at AD/PD: this may have been "too little, too late."

Where the Hope Lives

The negative EVOKE results do not close the door on the metabolic pathway — they redirect it.

Several research directions are actively being pursued in the wake of these trials.

First and most compelling is the question of prevention. Multiple experts have called for trials testing GLP-1 agonists not in patients with existing symptomatic Alzheimer's, but in older adults who are metabolically at risk and biomarker-positive but not yet symptomatic — targeting the fifteen-to-twenty year pre-clinical window when intervention may actually matter.

Second, the EVOKE biomarker data has raised serious interest in combination therapy. Anti-amyloid drugs like Leqembi and Kisunla slow cognitive decline by roughly 30%. That leaves 70% of the disease's progression unaddressed. The metabolic and neuroinflammatory pathways that GLP-1 agonists do appear to modulate represent exactly the kind of complementary targets that a combination approach would need. As one scientist framed it at CTAD: this is precisely how cancer treatment evolved — from single-drug regimens that partially worked to combination regimens that transformed survival. Alzheimer's will almost certainly follow the same arc.

Third, the broader GLP-1 class continues to be investigated. Semaglutide is one agent in one formulation. Other GLP-1 and dual GLP-1/GIP agonists with different CNS penetration profiles, dosing regimens, and mechanisms may behave differently. The metabolic pathway has not been disqualified — one specific drug at one specific dose in one specific patient population did not clear the bar.

The Alzheimer's drug pipeline currently contains 182 active clinical trials. More than 70% of them target non-amyloid pathways. The era of single-target, single-drug thinking is giving way to something more sophisticated, more personalized, and ultimately more likely to succeed.

Takeaways

The EVOKE and EVOKE+ trials were negative. Oral semaglutide did not slow cognitive or functional decline in patients with early symptomatic Alzheimer's disease across any primary or secondary clinical endpoint.

The biomarker data was meaningful but insufficient. Semaglutide produced approximately 10% reductions in CSF tau biomarkers and neuroinflammation markers — statistically real effects that did not translate into clinical benefit at the disease stage studied.

The prevention hypothesis remains alive. The observational data showing reduced dementia risk in diabetic patients on GLP-1 agonists studied a fundamentally different question — risk reduction, not disease modification — and that question has not been answered by EVOKE.

Combination therapy is the direction the field is moving. With anti-amyloid drugs addressing roughly 30% of decline, the remaining pathways — metabolic, inflammatory, vascular, synaptic — represent the frontier. GLP-1 biology may yet play a role in that combination landscape.

For patients with diabetes and obesity, nothing has changed. Semaglutide remains a highly effective, well-validated treatment for its approved indications. Managing metabolic health is still one of the most important things you can do for your brain — the EVOKE results do not alter that fundamental truth.

Citations and References

Cummings J, et al. Efficacy and safety of oral semaglutide 14 mg in early-stage symptomatic Alzheimer's disease (evoke and evoke+): two phase 3, randomised, placebo-controlled trials. The Lancet. 2026. https://doi.org/10.1016/S0140-6736(26)00459-9

Novo Nordisk. Evoke phase 3 trials did not demonstrate a statistically significant reduction in Alzheimer's disease progression. Press Release, November 24, 2025. https://www.biospace.com/press-releases/novo-nordisk-a-s-evoke-phase-3-trials-did-not-demonstrate-a-statistically-significant-reduction-in-alzheimers-disease-progression

Alzheimer's Drug Discovery Foundation. New Data from Semaglutide Trials Provides Critical Insights. December 4, 2025. https://www.alzdiscovery.org/news-room/announcements/new-data-from-semaglutide-trials-provides-critical-insights

Alzheimer Europe. Results of EVOKE and EVOKE+ trials show no effect of oral semaglutide on AD progression. December 2025. https://www.alzheimer-europe.org/news/results-evoke-and-evoke-trials-show-no-effect-oral-semaglutide-ad-progression

Clinical Trials Arena. AD/PD 2026: Novo Nordisk's semaglutide fails to demonstrate any benefit for early Alzheimer's. March 2026. https://www.clinicaltrialsarena.com/analyst-comment/ad-pd-2026-novo-nordisk-semaglutide-early-alzheimers/