Oral Semaglutide and Cognitive Decline Risk

A retrospective study of 246 elderly patients found fewer new diagnoses of cognitive impairment among those taking oral semaglutide than among those taking a DPP-4 inhibitor. A letter published in the same journal a month later responded to it. Neither paper is a randomized trial, and the finding sits next to a much larger, purpose-built Alzheimer's trial that found no cognitive benefit from the same drug. Here is what the numbers actually show, and what they don't.

What the underlying study found

The study behind this discussion, led by Armentaro and colleagues, followed 246 matched elderly patients with heart failure with preserved ejection fraction (HFpEF), obesity, and type 2 diabetes for one year^[1]. HFpEF means the heart's main pumping chamber still contracts with roughly normal strength, but doesn't relax and fill properly between beats, so blood backs up and produces the usual breathlessness and fluid retention of heart failure. Everyone in the cohort was taking either oral semaglutide (marketed as Rybelsus for type 2 diabetes) or a DPP-4 inhibitor, an older class of oral diabetes drugs.

Over the year of follow-up, the researchers recorded 106 new cases of cognitive impairment across the whole group^[1]. The split was uneven: 33 new cases in the group on oral semaglutide, versus 73 in the group on a DPP-4 inhibitor, a difference the authors reported as statistically significant. A logistic regression model estimated a 77% lower relative risk of incident cognitive impairment associated with oral semaglutide compared with DPP-4 inhibitors.

Why a comment followed, and what we could and couldn't verify

Five weeks after the original paper appeared, the same journal, Internal and Emergency Medicine, published a short letter from a different set of authors with the same title plus the word "comment"^[2]. Letters like this are a routine part of how medical journals work: another group reads a paper and writes in with a question, a caveat, or an alternative reading. The letter carries no public abstract and sits behind a subscription wall, so we could not independently verify what specific point it raises, and we are not going to guess. What we can say with confidence is what kind of paper it is: expert commentary responding to a single retrospective study, not new trial data of its own.

That distinction matters more than it sounds. A 246-patient, one-year, single-cohort retrospective comparison can show a real association, but it cannot rule out confounding: the two treatment groups may have differed in ways the matching didn't fully capture, such as how long each group had lived with diabetes, how well their blood sugar was controlled, or which patients their prescribers chose to switch onto a newer drug in the first place.

The bigger trial says something different

Oral semaglutide has also been tested directly against Alzheimer's disease in two large, purpose-built trials called EVOKE and EVOKE+, run by Novo Nordisk and published in The Lancet^[3]. These trials enrolled 3,808 people aged 55 to 85 with mild cognitive impairment or mild dementia and confirmed amyloid buildup in the brain, the biological hallmark of Alzheimer's disease, and followed them for two years on semaglutide or placebo.

The result was a clear miss on the primary measure. In the EVOKE trial, the change in a standard dementia-progression score (the Clinical Dementia Rating, Sum of Boxes) was nearly identical between semaglutide and placebo, and the difference was not statistically significant. EVOKE+ found the same pattern^[3]. Semaglutide did lower some blood markers of brain inflammation, but that biological signal did not translate into slower decline on the scale that actually measures a patient's day-to-day function. Novo Nordisk stopped the trials' planned one-year extension after the results came in^[4].

How two studies can both be right

These are not the same claim, and one does not cancel the other out. The retrospective HFpEF study measured a broad category, any newly diagnosed cognitive impairment, in older adults with heart failure, obesity and diabetes, most of whom did not have Alzheimer's disease. EVOKE and EVOKE+ measured progression on a specific dementia scale in people who already had confirmed early Alzheimer's disease. Different population, different outcome, different study design. It's entirely possible for a drug's cardiometabolic benefits, better blood sugar control, less inflammation, less vascular strain on the brain, to show up as fewer cognitive-impairment diagnoses in a mixed elderly population, while doing nothing measurable to slow an amyloid-driven disease that has already taken hold.

What this does not show

This is not evidence that semaglutide prevents dementia, and it is not a reason to start or switch a diabetes medication for brain-health purposes. The retrospective study is hypothesis-generating: a single-center or limited-cohort observational finding that needs to be tested in a randomized trial designed for that specific question before anyone can call it a preventive effect. We don't yet know whether a larger, longer, randomized comparison of oral semaglutide against DPP-4 inhibitors, specifically measuring cognitive outcomes in this HFpEF-and-obesity population, would confirm the 77% figure, shrink it, or erase it entirely.

If you are managing type 2 diabetes alongside heart failure or obesity and have questions about semaglutide or how it compares with other options, that conversation belongs with the clinician managing your care, not with a single observational paper.

Medical disclaimer: This article is for educational and journalistic purposes only and does not constitute medical advice. Semaglutide is a prescription medicine in every jurisdiction we cover. Always consult a qualified healthcare professional before starting, stopping, or changing any medication. PeptideMethods.com does not sell, distribute, or facilitate the sale of any peptide product.