A 32-week trial of slow-release exenatide in 32 patients with mild cognitive impairment found no benefit on cognitive performance, with a concerning signal of worsening in women.
No cognitive benefit (p=0.17)32 weeks of weekly exenatide injections produced no significant improvement on the ADAS-Cog11 cognitive test in patients with mild cognitive impairment, despite reducing blood sugar and weight
What the researchers found
The primary endpoint — change in ADAS-Cog11 cognitive score at 32 weeks — showed no significant difference between exenatide and control groups (p=0.17). A gender interaction was detected (p=0.04), driven by worsening ADAS-Cog11 scores in women randomized to exenatide (p=0.018), even after adjusting for age, education level, dysglycemia, and baseline cognitive scores.
Exenatide did demonstrate expected metabolic effects: fasting plasma glucose decreased (p=0.02) and body weight decreased (p=0.03) in the treatment group, confirming GLP-1 receptor engagement. However, these metabolic improvements did not translate into cognitive benefits.
Why it matters
There's enormous interest in whether GLP-1 drugs could prevent or treat Alzheimer's disease, fueled by promising animal data and observational studies. This rigorous proof-of-concept trial provides an important reality check: exenatide did not slow cognitive decline in mild cognitive impairment over 32 weeks. The concerning signal in women adds complexity. These results are crucial for tempering premature optimism and guiding the design of future trials with different drugs, doses, or patient populations.
How the study worked
Randomized proof-of-concept clinical trial (NCT03881371) with 32 patients (16 female) with mild cognitive impairment. Participants were randomized to slow-release exenatide (2 mg subcutaneous once weekly, n=17) or no treatment (n=15) for 32 weeks. The primary endpoint was change in ADAS-Cog11 score. Secondary endpoints included additional cognitive tests and plasma biomarkers of GLP-1 receptor engagement. Analysis was conducted by intention to treat.
What this study cannot tell us
Very small sample size (32 patients) severely limits statistical power to detect moderate cognitive effects. The trial used no treatment rather than placebo as the control, which does not account for placebo effects. The 32-week duration may be too short to detect neuroprotective effects in a slowly progressive condition. Exenatide may have limited brain penetration compared to newer GLP-1 drugs. The study focused on mild cognitive impairment broadly, not specifically on Alzheimer's disease pathology.
How to read the evidence
This is a registered, randomized proof-of-concept clinical trial, which is a reasonable study design for early investigation. However, the very small sample size (32 patients), open-label design (no placebo control), and single-site nature significantly limit the strength of both the negative finding and the gender interaction signal.
When this study was published
Published in 2024, this trial addresses one of the most active questions in neurology: whether GLP-1 drugs can protect against cognitive decline. Results from larger, placebo-controlled trials of semaglutide for Alzheimer's are expected to provide more definitive answers.
The bigger picture
The search for Alzheimer's treatments remains one of medicine's greatest challenges. GLP-1 receptor agonists were considered promising candidates based on neuroprotective effects in animals, but translating brain benefits from rodents to humans has proven difficult across many drug classes. This negative trial joins mixed results from other GLP-1/dementia studies, suggesting the relationship between GLP-1 signaling and neurodegeneration may be more complex than animal models indicate. Larger trials with newer, more brain-penetrant GLP-1 drugs like semaglutide are still ongoing.
Questions still open
- Would more brain-penetrant GLP-1 receptor agonists like semaglutide show different results in cognitive trials?
- Why did women specifically show cognitive worsening on exenatide — is this a real biological signal or a statistical artifact in a small sample?
- Would a longer treatment duration or earlier intervention (before cognitive impairment) change the outcome?
Common questions
Does this mean GLP-1 drugs can't help with Alzheimer's?
Why did the drug seem to worsen cognition in women?
Read the original research
Long-acting exenatide does not prevent cognitive decline in mild cognitive impairment: a proof-of-concept clinical trial.
Journal of endocrinological investigation, 47(9), 2339-2349
Citation
Dei Cas, A; Micheli, M M; Aldigeri, R; Gardini, S; Ferrari-Pellegrini, F; Perini, M; Messa, G; Antonini, M; Spigoni, V; Cinquegrani, G; Vazzana, A; Moretti, V; Caffarra, P; Bonadonna, R C. (2024). Long-acting exenatide does not prevent cognitive decline in mild cognitive impairment: a proof-of-concept clinical trial.. Journal of endocrinological investigation, 47(9), 2339-2349. https://doi.org/10.1007/s40618-024-02320-7