Semaglutide slows blood protein signature linked to future dementia risk

By Dr. Sanchari Sinha Dutta, Ph.D.Reviewed by Lauren HardakerAug 10 2026

In older adults with cardiovascular disease and overweight or obesity, semaglutide slowed worsening of a blood-based dementia risk signature, offering new clues about how cardiometabolic treatment may influence biological pathways linked to brain health.

Study: Semaglutide attenuates a proteomics-based dementia risk signature in older adults with overweight or obesity and cardiovascular disease without diabetes: A post hoc analysis of the SELECT phase 3 trial. Image Credit: Lightspring / Shutterstock

Semaglutide, a glucagon-like peptide-1 (GLP-1) receptor agonist used to treat diabetes and obesity, attenuated worsening of a blood-based proteomic signature associated with future dementia risk in overweight or obese older adults with cardiovascular disease without diabetes, according to a new study published in the journal Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring.

Background

Dementia is a highly prevalent age-related disorder that can arise from multiple interacting pathological processes, including neurodegenerative proteinopathies, cerebrovascular disease, and neuroinflammation. Alzheimer’s disease is the most common cause of dementia. The condition is associated with progressive memory loss and cognitive decline.

Alzheimer’s disease-specific biomarkers, such as amyloid-beta and tau, measured in the cerebrospinal fluid or plasma have significantly improved disease diagnosis. However, these biomarkers cannot fully capture the cumulative impact of systemic complications, including cardiovascular and metabolic diseases, on the all-cause dementia risk.

Plasma proteomics, involving multi-protein panels and machine learning, has shown promise in predicting future dementia risk. The Dementia SomaSignal Test (dSST) is a machine learning-derived blood plasma risk score based on 25 specific proteins that predicts individualized 5-year and 20-year risk of an all-cause dementia diagnosis.

Given the consistent association between cardiometabolic risk factors and future dementia risk, researchers from Novo Nordisk (the manufacturer of semaglutide), Amsterdam UMC, and Illumina conducted a statistical analysis of the Heart Disease and Stroke in Patients With Overweight or Obesity (SELECT) trial findings to evaluate the effect of semaglutide on the circulating proteomic signatures associated with future dementia risk.

Study design

The SELECT trial was a randomized, double-blind, placebo-controlled study investigating the effect of the glucagon-like peptide-1 receptor agonist semaglutide on cardiovascular outcomes in 17,604 obese or overweight adults aged 45 years or older with established atherosclerotic cardiovascular disease without diabetes.

The current analysis included data from 2970 SELECT trial participants aged 65 years or above who were randomized to receive either once-weekly semaglutide, titrated to a target dose of 2.4 milligrams over 16 weeks, or placebo. Serum samples collected at baseline and week 104 were analyzed using the dSST to predict 5-year and 20-year dementia risk.

Key findings

The study analysis revealed an increase in proteomics-predicted 5-year dementia risk in both the semaglutide and placebo (control) groups over the follow-up period of 104 weeks. However, the magnitude of increase was 2.5-fold lower in participants who received semaglutide. This corresponded to a 26.0% lower predicted dementia event rate in these participants.

Similarly, the magnitude of the increase in proteomics-predicted 20-year dementia risk was 1.67-fold lower in the semaglutide group, corresponding to an 8.8% lower predicted dementia event rate.

These findings indicate a more modest effect of semaglutide on the 20-year dementia risk signature compared to the 5-year risk signature. The authors suggested that this difference could reflect greater short-term sensitivity of the 5-year model, more reversible near-term inflammatory or vascular processes, or differences in statistical calibration between the prediction horizons. The 104-week intervention may also have greater effects on near-term pathways, whereas long-horizon risk reflects accumulated lifetime exposures that may be less amenable to relatively short-term intervention.

The analysis of the distribution of dSST risk categories at week 104 revealed a significant downward shift in the semaglutide group. Specifically, semaglutide was associated with 36% lower odds of being classified into a higher dementia risk category compared to placebo.

Among 18 participants experiencing dementia-related adverse events, baseline dSST risk categories were either high or medium-to-high. None of these participants was classified as low risk at baseline. However, these events were not systematically adjudicated, and the small number of cases and reliance on adverse-event reporting mean the findings should be interpreted cautiously.

Study significance

The study findings suggest that 104 weeks of treatment with semaglutide may modify systemic proteomic signatures associated with long-term dementia risk in a cardiometabolic population. Overall, the study findings are consistent with the possibility that semaglutide influences dementia-related biology through metabolic, inflammatory, vascular, and potentially neurological pathways, although the mechanisms responsible were not directly established.

About 76% of dementia-associated proteins in the dSST have established roles in cardiometabolic disease, suggesting that this protein signature may reflect the benefits of semaglutide on the shared mechanisms linking cardiometabolic dysfunction to dementia risk.

Mechanistic evidence suggests that glucagon-like peptide-1 receptor agonists, including semaglutide, may influence biological processes relevant to dementia by improving metabolic regulation, reducing adiposity and systemic and neuro-inflammation, and improving endothelial function and vascular health.

Real-world studies have linked glucagon-like peptide-1 receptor agonist-based treatment to lower clinical diagnosis of Alzheimer’s disease and dementia in patients with type 2 diabetes, although confounding and channeling bias remain concerns.

Notably, the study finds a 28% attenuation of the treatment effect on the 20-year predicted dementia-risk signature after adjusting for changes in body mass index (BMI) during the study period. The persistence of approximately 72% of the treatment association after this adjustment is consistent with the possibility that factors beyond weight reduction may be associated with the observed proteomic signature modification in the semaglutide group.

The study analyzed data from the SELECT trial, which was designed to evaluate cardiovascular events but lacked optimized patient selection and data-collection procedures to track cognitive health. The proteomic analysis was also post hoc, and no prespecified multiple-testing adjustment was applied to these endpoints.

Moreover, the study relied on a biomarker-based predicted risk score rather than standardized cognitive test scores or adjudicated dementia outcomes. The dSST was also developed and validated using ethylenediaminetetraacetic acid (EDTA)- plasma samples, whereas the SELECT analysis used serum, although the researchers found high concordance between plasma- and serum-derived scores. These limitations restrict any causal interpretation that semaglutide-induced changes in the proteomic signature translate into a lower clinical risk of developing dementia.

Future studies should investigate whether proteomics-predicted risk reduction by semaglutide translates to slower cognitive decline or lower dementia incidence in individuals with cardiometabolic disease. The study was funded by Novo Nordisk, which participated with academic collaborators in study design, data collection, analysis, and preparation and review of the manuscript. Two authors were Novo Nordisk employees and minor shareholders.