Semaglutide, the blockbuster weight-loss and diabetes drug marketed as Ozempic and Wegovy, has failed to slow cognitive decline in people with Alzheimer’s disease, according to Phase 3 trial results announced in November 2025 and presented in full at the AD/PD 2026 International Conference on March 19, 2026. The two identical trials, called EVOKE and EVOKE+, enrolled 3,808 participants aged 55 to 85 years with mild cognitive impairment or mild dementia caused by Alzheimer’s disease. Despite earlier hopes that semaglutide might protect the brain through weight loss or anti-inflammatory pathways, neither trial showed meaningful slowing of disease progression when compared to placebo. In the EVOKE trial, patients on semaglutide declined 2.2 points on the Clinical Dementia Rating Scale–Sum of Boxes (CDR-SB) over 104 weeks, versus 2.2 points in the placebo group—a difference of 0.06 points that was statistically insignificant (P = .7727).
The EVOKE+ trial showed a similarly small difference: semaglutide patients declined 2.1 points compared to 2.0 in the placebo group, a 0.15-point difference also lacking statistical significance (P = .4604). This outcome marks a significant disappointment for researchers and patients hoping that repurposing an existing medication could offer a new avenue for slowing Alzheimer’s progression. Unlike some recent Alzheimer’s drugs that target amyloid plaques—the misfolded proteins that accumulate in the brains of people with the disease—semaglutide was believed to work through a different mechanism: by reducing inflammation, supporting metabolic health, and potentially protecting brain cells from damage. The trials required participants to have confirmed amyloid positivity, meaning they had measurable amyloid in their brains at enrollment. Yet even in this carefully selected population, the drug’s benefits, if any, proved too small to matter clinically.
Table of Contents
- Why Researchers Expected Semaglutide to Help Alzheimer’s Disease
- The Trial Design and What the Numbers Reveal
- The Puzzling Discrepancy Between Biomarkers and Clinical Outcomes
- What This Means for Patients, Caregivers, and Treatment Decisions
- Safety Profile and Tolerability Findings
- How Semaglutide Fits into the Broader Alzheimer’s Research Landscape
- What the Trial Results Mean for Prevention and Lifestyle
Why Researchers Expected Semaglutide to Help Alzheimer’s Disease
The idea that semaglutide might slow Alzheimer’s progression emerged from several converging lines of evidence. GLP-1 receptor agonists—the drug class to which semaglutide belongs—have been shown in laboratory and animal studies to reduce neuroinflammation, protect neurons from damage, and improve metabolic markers associated with brain health. There is also epidemiological evidence linking weight loss and metabolic health improvements to better cognitive outcomes in older adults. Some early observational research suggested that people taking GLP-1 agonists for diabetes or weight loss had lower rates of neurodegenerative disease, though these studies could not prove causation.
The hope was that semaglutide might work synergistically with amyloid pathology reduction, building on the modest benefits seen with monoclonal antibodies like lecanemab, which targets amyloid but has a clinically small effect size. Additionally, semaglutide has a substantial track record in clinical trials and millions of real-world users, lending confidence in its tolerability and safety profile. Unlike drugs being developed from scratch specifically for Alzheimer’s disease, repurposing an established medication can accelerate drug development and reduce uncertainty. The drug’s widespread use also meant that patient recruitment for the EVOKE trials was feasible, with 3,808 participants completing 104 weeks of treatment or placebo with follow-up data through an additional 52-week extension phase.
The Trial Design and What the Numbers Reveal
The EVOKE and EVOKE+ trials were 104-week, double-blind, randomized controlled studies in which participants received either semaglutide or placebo injections once weekly. Both trials required participants to have confirmed amyloid positivity at baseline, using positron emission tomography (PET) imaging or cerebrospinal fluid biomarkers, ensuring that all enrolled participants had measurable Alzheimer’s pathology. This was a deliberate design choice: by focusing on amyloid-positive individuals with mild cognitive impairment or mild dementia, the researchers hoped to catch disease progression early enough that intervention might make a measurable difference. The primary outcome measure was the change from baseline to week 104 on the CDR-SB, a 18-point scale tracking cognitive and functional decline that clinicians commonly use in Alzheimer’s trials. The results in both trials were unambiguous: semaglutide did not slow cognitive decline compared to placebo.
The CDR-SB differences—0.06 points in EVOKE and 0.15 points in EVOKE+—are statistically nonsignificant and clinically negligible. A clinically meaningful slowing of cognitive decline is typically considered to be a difference of at least 0.5 points on the CDR-SB, a threshold used in regulatory decisions and clinical practice. Neither trial came close to this threshold. The trials were powered to detect smaller differences than historically seen with other Alzheimer’s drugs, meaning the studies had sufficient sample size to catch even modest benefits if they existed. The fact that semaglutide did not outperform placebo in either trial, despite enrolling thousands of participants, strengthens the conclusion that any real benefit, if present, is negligible.
The Puzzling Discrepancy Between Biomarkers and Clinical Outcomes
One of the most striking findings in the EVOKE and EVOKE+ trials was an apparent paradox: semaglutide improved several Alzheimer’s-related biomarkers—the biological markers measured in blood, cerebrospinal fluid, or brain imaging that reflect disease pathology—yet this improvement did not translate into slowing clinical symptoms or cognitive decline. Biomarkers such as phosphorylated tau, amyloid levels, and neurofilament light chain all showed improvement in the semaglutide groups compared to placebo. This observation raises a fundamental question about what these biomarkers actually predict about clinical benefit.
This disconnect between biomarker improvement and clinical outcome is not unique to semaglutide but illustrates a perennial challenge in Alzheimer’s drug development. Biomarkers can shift in response to an intervention, but they do not always correlate with whether a patient notices or experiences any slowing of their cognitive decline or functional abilities. The improvement in biomarkers might reflect a genuine biological effect—perhaps semaglutide truly does reduce inflammation or support neuronal health—but if that effect is too small or occurs too late in the disease process, it may not prevent the behavioral, cognitive, and functional symptoms that matter most to patients and caregivers. For families struggling with Alzheimer’s disease, a drug that improves biomarkers while memory continues to decline at the same rate offers little comfort or practical value.
What This Means for Patients, Caregivers, and Treatment Decisions
For people with mild cognitive impairment or mild dementia due to Alzheimer’s disease who were hoping semaglutide might offer a new treatment option, these trial results represent a setback. Currently, the disease-modifying treatments available for Alzheimer’s disease include lecanemab (approved by the FDA and recommended for people with mild cognitive impairment or mild dementia with confirmed amyloid pathology) and donanemab (a similar amyloid-targeting monoclonal antibody in late-stage development). Both of these drugs show modest slowing of cognitive decline—on the order of 35 percent slowing over 18 months—compared to placebo. Semaglutide, by contrast, showed no benefit.
Patients and clinicians considering treatment options should not view semaglutide as an alternative to established Alzheimer’s disease treatments. However, semaglutide may still be appropriate for people with Alzheimer’s disease who have comorbid obesity or type 2 diabetes, conditions that are common and that the drug does effectively treat. A person with early-stage Alzheimer’s disease and obesity might reasonably take semaglutide for weight loss and metabolic health, but the expectation should be that the semaglutide itself will not slow cognitive decline. Conversely, a person with Alzheimer’s disease for whom lecanemab is indicated should pursue that treatment based on the current evidence, regardless of whether they are also overweight. Combining multiple disease-modifying approaches—such as taking both lecanemab and semaglutide—has not been studied and should not be assumed to be additive or beneficial without evidence.
Safety Profile and Tolerability Findings
The EVOKE and EVOKE+ trials confirmed that semaglutide was well-tolerated in this population of older adults with cognitive impairment, with a consistent safety profile matching what has been observed in prior trials of the drug for weight loss and diabetes. The most common side effects associated with semaglutide remain gastrointestinal in nature—nausea, vomiting, diarrhea, and constipation—effects that typically emerge in the early weeks of treatment and often improve over time as the body adjusts. For people with Alzheimer’s disease, who may already struggle with appetite, weight loss, or nutritional intake, these gastrointestinal effects merit careful consideration.
A person with advanced cognitive decline might not be able to communicate that they are experiencing nausea or other adverse effects, making close monitoring by family or caregivers essential. One important decision made following the trial results was to discontinue the planned 52-week extension phase in both EVOKE and EVOKE+ trials. This discontinuation was based on the lack of benefit observed in the primary 104-week treatment phase; continuing the extension would have exposed participants to an additional year of treatment for no demonstrated clinical benefit. This decision reflects a principle of research ethics: once a treatment has been shown not to work, continuing to give it to research participants is difficult to justify.
How Semaglutide Fits into the Broader Alzheimer’s Research Landscape
The failure of semaglutide to slow cognitive decline in Alzheimer’s disease suggests that targeting inflammation and metabolic pathways alone may not be sufficient to meaningfully slow the neurodegenerative process in people who already have symptoms. This does not mean that inflammation and metabolism are irrelevant to Alzheimer’s disease—they clearly play roles in brain health and may influence disease risk—but it suggests that once amyloid and tau pathology are already established, these other mechanisms may be secondary drivers of ongoing cognitive decline.
The field has gradually shifted toward targeting amyloid and tau directly through monoclonal antibodies and tau-targeting drugs in clinical development, a strategy that has yielded at least modest clinical benefits. Future research may explore combining semaglutide or other GLP-1 agonists with amyloid-targeting therapies in people with mild cognitive impairment or mild dementia, but such combination trials would need to be designed and conducted before any such approach could be recommended. It is also possible that GLP-1 agonists might be beneficial in preventing the transition from normal cognition to mild cognitive impairment, or in slowing decline in people with more advanced dementia, but these questions remain unanswered.
What the Trial Results Mean for Prevention and Lifestyle
While semaglutide did not slow cognitive decline in people with established Alzheimer’s disease, the question of whether weight management and metabolic optimization might prevent or delay the onset of cognitive impairment in people without symptoms remains open. Observational studies have long suggested that cardiovascular risk factors, obesity, and metabolic dysfunction in midlife are associated with higher risk of cognitive impairment and dementia later in life. It is possible that maintaining a healthy weight and good metabolic control in one’s 40s and 50s could reduce the risk of developing cognitive symptoms decades later. However, such effects are generally modest and not yet definitively proven, and no drug has been shown to prevent dementia in cognitively normal older adults.
Lifestyle interventions—regular physical activity, cognitive engagement, social connection, quality sleep, and a heart-healthy diet—remain the most evidence-based approaches to cognitive health for people without symptoms. The EVOKE and EVOKE+ trials collectively enrolled thousands of participants and followed them for 104 weeks of treatment. The topline results were announced in November 2025, and the full datasets were presented at the AD/PD 2026 International Conference on March 19, 2026, allowing the scientific community to scrutinize the detailed findings. These large, well-designed trials represent a significant investment of time, resources, and participant effort—and an honest null result, while disappointing, is an important contribution to medical knowledge that helps prevent future patients from being given ineffective treatments.
- —
