GLP-1, Brain and Behavior: Alcohol, Alzheimer's and Honest Science

The GLP-1 receptor is not alone in the pancreas. It is also expressed in key brain areas of the reward system, appetite regulation and behavior modulation. When that biological piece began to be understood, researchers asked the obvious: if GLP-1 modulates the desire for food, could it modulate other desires?

That question opened a research agenda that in recent years has given surprising results in some territories and disappointing results in others. In 2025, the first randomized controlled trial in humans showed that semaglutide reduced alcohol craving and consumption in people with alcohol use disorder. Around the same time, two large-scale phase 3 trials in Alzheimer's and Parkinson's ended in negative results.

These two stories, read together, are what honest science says today about GLP-1 and the brain. In this guide we go over what works, what doesn't work, and what remains to be seen.

The Biological Promise: Why Was It Tested in the Brain?

The GLP-1 receptor is present in multiple brain areas with known function:

• Hypothalamus:
• Area Postema:
• Hippocampus:
• Ventral tegmental area (VTA):
• Nucleus accumbens (NAc):

When semaglutide activates receptors in the VTA and NAc, it modulates the dopaminergic response to pleasurable stimuli. This explains the effect on appetite for food — and opened the hypothesis that it could modulate other types of "desire."

Additionally, Wong C.K., Drucker D.J. and collaborators published in in 2024 decisive evidence: the systemic anti-inflammatory effect of GLP-1 is exerted predominantly via the central GLP-1 receptor, with adrenergic and opioid signaling. Low-grade neuroinflammation is a plausible mechanism in neurodegenerative diseases such as Alzheimer's and Parkinson's. Therefore, the researchers decided to test GLP-1 in these two areas.

The hypothesis was reasonable. As we will see, promising biology did not guarantee a clinical outcome.

Hendershot 2025: The First RCT of Semaglutide in Alcohol

Hendershot C.S., Bremmer M.P., Klein K.R. et al published in JAMA Psychiatry 2025 (volume 82, issue 4, pages 395–405) the first randomized controlled trial of semaglutide in people with alcohol use disorder (AUD).

Características clave del estudio:

The study is small (48 people) and short (9 weeks), but it is the first causal demonstration in humans. Before Hendershot 2025, there was only evidence in animal models and observational cohorts. Now there is a controlled positive signal — which is very different.

The Cohorts That Reinforce the Signal

Two independent cohorts have shown patterns consistent with Hendershot's finding, on much larger scales:

Wang W., Volkow N.D. and collaborators — Nature Communications 2024

Volumen 15, artículo 4548.

TriNetX cohort, millions of clinical records. Semaglutide users showed lower incidence and recurrence of AUD than comparators with other antidiabetics. That the finding comes from the laboratory of Nora Volkow (director of the NIH NIDA, a world authority on addictions) adds interpretive weight.

Lähteenvuo M. et al. — JAMA Psychiatry 2025

Cohort from Nordic records using design: the same person is compared with themselves in periods of use vs non-use of GLP-1. This design reduces confounding by individual characteristics. Result: lower risk of hospitalization for AUD during periods of GLP-1 use versus periods of no use.

Three lines of evidence converging — a small but causal phase 2 RCT, a US cohort of millions, and a Nordic cohort with robust methodology. It is the convergence that gives weight, not each isolated study.

The Mechanism: Modulation of the Reward System

What exactly does semaglutide do in the addicted brain? The best supported hypothesis points to the mesolimbic dopaminergic system.

Alcohol, like other addictive substances, activates the release of dopamine in the nucleus accumbens via projections from the VTA. That dopaminergic signal is what the brain registers as a "reward" and what motivates repetition of the behavior. GLP-1 receptors in the VTA and NAc modulate that dopaminergic response — they attenuate it, in simple terms. The behavioral result: less craving, less reinforcement, less consumption.

Wong and Drucker's work in Cell Metabolism 2024 adds an additional layer: activation of the central GLP-1 receptor engages κ-opioid and α1-adrenergic pathways. The κ-opioid system is involved in states of dysphoria and in the negative component of addiction (the discomfort that motivates using again). If GLP-1 also modulates this axis, it could act both on the search for reward and on the relief of discomfort — two classic arms of addiction.

Beyond Alcohol: Signs in Other Behaviors

The smoking subgroup in Hendershot 2025 showed reduction in cigarette smoking. Observational cohorts suggest reduction in opioid craving and compulsive behaviors (such as impulsive buying or gambling). If the reward system is the same, the modulating effect could generalize.

Quddos F. and collaborators published in 2023 (volume 13, article 20998) data on semaglutide and tirzepatide in people with obesity who also consumed alcohol: both peptides were associated with reduced alcohol consumption. The discovery was not intended as a primary objective — it emerged as a side observation. Sometimes lateral finds are the most informative.

Important Warning:

EVOKE and EVOKE+: Negative Result in Alzheimer's

If the previous story was one of promise and partial confirmation, this one is one of promise and disappointment.

Cummings J., Scheltens P. and collaborators published the results of the EVOKE and EVOKE+ trials in 2026. They were designed to answer a clear question: can oral semaglutide 14 mg slow the clinical progression of early-stage Alzheimer's?

→ Diseño robusto

3,808 participants (55–85 years) with mild cognitive impairment or mild Alzheimer's dementia. Oral semaglutide 14 mg vs placebo for 104 weeks. 566 sites in 40 countries. Sufficient statistical power to detect clinically relevant effects.

→ Resultado primario negativo

In EVOKE, the difference in CDR-SB (Clinical Dementia Rating–Sum of Boxes, standard scale) between semaglutide and placebo was −0.06 points (95% CI −0.48 to 0.36; P=0.77). In EVOKE+, the difference was 0.15 points (95% CI −0.24 to 0.54; P=0.46). No clinically significant difference, no delay in time to dementia.

→ The paradox of the CSF substudy

The cerebrospinal fluid substudy showed that semaglutide reduced several pathology markers (p-tau181, p-tau217, np-tau205, YKL-40, total tau, neurogranin) by up to 10%. That is to say: the biomarkers improved, but this biochemical improvement did not translate into a measurable clinical benefit.

The company discontinued the extension period. EVOKE is one of the great negative trials of recent years in dementia. And it's a powerful lesson: moving a biomarker is not equivalent to moving a life.

Exenatide-PD3: Negative Result in Parkinson's

Vijiaratnam N., Foltynie T. and collaborators published the results of Exenatide-PD3 in 2025. Context matters: in 2017, the same Foltynie group published a phase 2 trial (Athauda et al.) that had shown a positive signal with exenatide in Parkinson's. That signal motivated the confirmatory phase 3 trial.

Características del fase 3:

• 194 people with moderate Parkinson's.
• Exenatida QW (semanal) 2 mg vs placebo durante 96 semanas.
• Diseño doble ciego, controlado.

The result: no significant difference in the MDS-UPDRS-3 motor scale in the OFF state between the active group and the placebo. The positive signal from phase 2 was not replicated in phase 3.

Other trials in Parkinson's have had mixed results. NLY01 (pegylated exenatide), published by McGarry A. and collaborators in 2024, was also negative in the primary endpoint, although it showed a signal in the subgroup under 60 years of age. LixiPark (Meissner W.G. et al., NEJM 2024) with lixisenatide did show less motor progression in early Parkinson's, but with limited gastrointestinal tolerance. The outlook: some signals in subgroups, but no positive large phase 3 RCTs.

The Paradox: When Biomarkers Improve Without Clinical Change

EVOKE was not the only trial where biomarkers moved in the expected direction without clinical translation. It is a phenomenon that neurology has been seeing with some frequency, especially in dementia. What does it mean?

Varias explicaciones posibles:

In EVOKE, people who already had cognitive impairment were treated. Perhaps the neuronal damage was already irreversible and moving biomarkers could not reverse the trajectory. The precautionary hypothesis: start earlier, in people at risk but without symptoms.

Oral semaglutide has limited bioavailability. The effective brain dose may have been insufficient. More potent formulations or more efficient routes could change the result.

Perhaps a subgroup of Alzheimer's (defined by age, genetics, specific biomarkers) does benefit, and the signal is diluted in the overall average. Posthoc analyzes are ongoing.

It is possible that neuroinflammation is not a decisive mechanism in clinical Alzheimer's, although it does affect biomarkers. The complexity of Alzheimer's has defeated many reasonable hypotheses.

Whatever explains the pattern, EVOKE reminds us of a central truth of evidence-based medicine: what seems mechanically reasonable does not always translate into clinical benefit. Biomarkers are imperfect substitutes. Only the outcomes that matter to the patient (function, quality of life, mortality) close the question.

What These Results Teach

Three lessons emerge from this combination of findings:

→ Efficacy is indication-specific

Just because semaglutide works to reduce cardiovascular events (SELECT), protect the kidney (FLOW) and improve the liver (ESSENCE) does not mean that it works for everything. Each indication is validated or discarded independently. The molecule is not universal — its effectiveness depends on the biological and clinical context.

→ Negative essays are as valuable as positive ones

Without EVOKE, thousands of people with Alzheimer's could be paying for a therapy that doesn't work. Without Exenatide-PD3, the same in Parkinson's. These essays close a chapter and open more precise questions. Science advances both when it discards and when it confirms.

→ The signal in alcohol and behavior is real but early

Three lines of converging evidence (phase 2 RCTs + 2 independent cohorts) build a promising body of data on GLP-1 and alcohol use disorders. But it is not yet an approved indication, and phase 3 trials will need to confirm before any regulatory authority considers authorizing it. Psychotherapy, approved medications, and support remain the standard of care in AUD.

Conclusion: an Honest Look

GLP-1 — semaglutide, tirzepatide, exenatide and others — modulate brain circuits related to appetite, reward and inflammation. That biology is real and opened up legitimate expectations in multiple areas.

What the 2024–2026 evidence suggests:

• In alcohol and possibly other cravings, the signal is positive and convergent between a phase 2 RCT and two large cohorts. Real promise, pending confirmation.
• In Alzheimer's and Parkinson's, the large phase 3 trials have been negative. Biomarkers improved in some cases, but this did not translate into clinical benefit. The original hypothesis — neuroprotection through anti-inflammation — was weakened by the findings, although not completely ruled out.

For those who observe this class of drugs with sober expectations, the message is clear: neither universal promise nor total rejection. Each indication has its own evidence, its own limits, and its own clinical conversation. What semaglutide does for the heart and kidney is very different from what it may or may not do for the aging brain — and understanding that distinction is what separates science from marketing.

Encuentra más information aquí:

Referencias Clave

1. Hendershot C.S., Bremmer M.P., Klein K.R. et al. Once-Weekly Semaglutide in Adults With Alcohol Use Disorder: A Randomized Clinical Trial. JAMA Psychiatry 2025;82(4):395–405. [Link]
2. Wang W., Volkow N.D., Berger N.A., Xu R. Associations of semaglutide with incidence and recurrence of alcohol use disorder in real-world cohorts. Nature Communications 2024;15:4548. [Link]
3. Lähteenvuo M. et al. GLP-1 Receptor Agonists and Hospitalization for Alcohol Use Disorder in Nordic Registries. JAMA Psychiatry 2025.
4. Quddos F. et al. Semaglutide and tirzepatide reduce alcohol consumption in individuals with obesity. Scientific Reports 2023;13:20998. [Link]
5. Cummings J., Scheltens P. et al. Efficacy and safety of oral semaglutide 14 mg in early-stage symptomatic Alzheimer's disease (EVOKE and EVOKE+). The Lancet 2026.
6. Vijiaratnam N., Foltynie T. et al. Exenatide once weekly for the treatment of moderate Parkinson's disease (Exenatide-PD3). The Lancet 2025.
7. McGarry A. et al. NLY01 (pegylated exenatide) in early Parkinson's disease. Lancet Neurology 2024;23(1):37–45. [Link]
8. Meissner W.G. et al. Trial of Lixisenatide in Early Parkinson's Disease (LixiPark). New England Journal of Medicine 2024;390:1176–1185. [Link]
9. Wong C.K., McLean B.A., Baggio L.L., Drucker D.J. et al. Central GLP-1 receptor activation mediates the systemic anti-inflammatory effect of GLP-1 receptor agonists. Cell Metabolism 2024;36(1):130–143. [Link]