GLP-1 Receptor Agonists in Addiction Psychiatry—Neurobiological Rationale, Emerging Clinical Evidence, and Cautions for Practice: A Narrative Review

Abstract

Glucagon-like peptide-1 (GLP-1) receptor agonists, originally developed for type 2 diabetes and obesity, have recently attracted interest as potential modulators of addictive behavior. This narrative review summarizes current knowledge on the neurobiological basis, randomized controlled trials, and psychiatric relevance of GLP-1 analogs in substance use disorders. English-language articles available at the time of the search were reviewed between February and April 2026, with emphasis on topics most relevant to psychiatric practice. The literature suggests that GLP-1 signaling influences reward processing, cue reactivity, stress responses, relapse vulnerability, and executive control through actions in the gut–brain axis and mesocorticolimbic circuitry. Early clinical findings are most encouraging in alcohol-related outcomes, including reductions in alcohol cue reactivity, craving, alcohol self-administration, and some measures of heavy drinking, whereas evidence in nicotine dependence is mixed and appears more consistent for limiting post-cessation weight gain than for improving abstinence itself. Evidence for other substance use disorders remains preliminary. Across randomized controlled trials, interpretation is limited by small sample sizes, short follow-up, heterogeneous endpoints, and selective populations. In addition, psychiatric and behavioral safety requires careful attention, particularly regarding rapid weight loss, excessive appetite suppression, restrictive eating, dehydration, and psychological destabilization in vulnerable individuals. At present, GLP-1 receptor agonists should be regarded as promising but unproven adjunctive candidates in addiction psychiatry, warranting further rigorous trials, structured monitoring, and interdisciplinary collaboration.

1. Introduction

Glucagon-like peptide-1 (GLP-1) is a pleiotropic incretin hormone derived from proglucagon, produced predominantly in intestinal enteroendocrine L-cells and in a population of brainstem neurons, and involved in a broad range of metabolic and central regulatory processes [1]. Although GLP-1 was first recognized for its glucose-dependent insulinotropic effects, subsequent work established that its physiological actions extend well beyond glycemic control and include slowing of gastric emptying, regulation of appetite and food intake, and wider effects on fluid balance, inflammation, and neural function [1]. On this basis, GLP-1 receptor agonists were developed initially for the treatment of type 2 diabetes and later became a major therapeutic advance in obesity care because of their clinically meaningful effects on body weight and eating behavior [1,2,3,4].

In parallel with their widespread clinical use, GLP-1 receptor agonists have drawn increasing attention outside diabetology and obesity medicine. A growing number of clinical observations, patient reports, and early translational studies have raised the possibility that these agents may influence craving, reinforcement, and compulsive patterns of consumption more broadly, including those related to alcohol, nicotine, and other substances [2,3,4]. This emerging interest is particularly relevant for psychiatrists, because it places GLP-1 receptor agonists at the intersection of metabolic and mental health, suggesting that pathways involved in appetite regulation may also overlap with mechanisms implicated in addiction and other compulsive behaviors [2,4].

This conceptual shift has been reinforced by repeated reports that patients treated with GLP-1 receptor agonists often describe not only a reduction in appetite and “food noise,” but also a lessening of persistent urges to use alcohol, tobacco, and other rewarding substances [2]. Such observations have encouraged a broader reconsideration of whether obesity and addiction, traditionally approached as distinct clinical entities, may share partially overlapping domains of craving, reward sensitivity, and compulsive consumption [2,4]. In this view, the therapeutic effects of GLP-1 receptor agonists may reflect more than metabolic modulation alone and may have implications for behavioral and psychiatric symptoms that are organized around dysregulated reward processing [1,2,4].

At the same time, the current state of the field calls for balance and scientific caution. The available literature consistently portrays GLP-1 receptor agonists as promising candidates in the addiction space, but has not yet established treatments for substance use disorders [2,3,4]. Early findings from preclinical models, clinical observations, and real-world analyses are encouraging, yet the evidence remains incomplete and heterogeneous, and important questions persist regarding efficacy, safety, durability of effect, and appropriate patient selection [2,3,4]. For psychiatric practice, the relevance of GLP-1 receptor agonists therefore lies not only in their possible future therapeutic applications, but also in the broader theoretical implications they carry for how compulsive overeating, substance use, and reward-driven behaviors are understood within a shared clinical framework [2,4].

Accordingly, GLP-1 receptor agonists should presently be viewed as established metabolic therapies with emerging psychiatric and addiction-related significance [1,2,3,4]. Their growing prominence invites closer dialog between endocrinology, obesity medicine, and psychiatry, and supports the need for a careful synthesis of what is currently known, what remains uncertain, and how these agents may eventually be positioned within addiction medicine [2,3,4]. This manuscript was developed as a clinically oriented narrative review of the literature concerning glucagon-like peptide-1 receptor agonists (GLP-1RAs) in addiction psychiatry. The objective was to synthesize current evidence on neurobiological mechanisms, translational relevance, emerging randomized controlled trials data, and safety considerations that may be pertinent to psychiatric and addiction practice. Although several recent reviews have addressed GLP-1 receptor agonists in substance use disorders, most have focused primarily on pharmacological mechanisms, preclinical evidence, metabolic outcomes, or broad addiction-related endpoints. The present review differs by adopting a clinically oriented psychiatric perspective. Its aim is not to duplicate existing systematic or pharmacological reviews, but to translate the rapidly evolving evidence into questions directly relevant to psychiatrists and addiction clinicians, including interpretation of randomized trial findings, patient selection, psychiatric and behavioral safety, monitoring of eating behavior and mood, and the risk of premature off-label use. In this sense, the novelty of the article lies in its practice-oriented synthesis of GLP-1 receptor agonists as promising but still unproven neurometabolic modulators at the interface of addiction, reward processing, metabolic health, and psychiatric vulnerability.

2. Materials and Methods

A structured literature search was performed in PubMed/MEDLINE in February 2026, with a final update conducted in April 2026, before submission. To improve coverage, additional articles were identified through manual review of the reference lists of relevant original studies, reviews, and clinical trial reports.

The search strategy used combinations of terms related to GLP-1 biology and addictive disorders. The search string included the following terms and their combinations: addictions, GLP-1, randomized controlled trials, neurobiology, pharmacology, and substance use disorder. The search period was not restricted.

Articles were considered for inclusion if they met the following criteria: (1) published in English; (2) indexed in PubMed/MEDLINE; and (3) relevant to at least one of the predefined review domains of neurobiology of GLP-1 signaling in reward or addiction-related circuits, preclinical evidence in substance-related models, human experimental or clinical studies involving GLP-1RAs in substance use disorders, or psychiatric/behavioral safety issues associated with GLP-1RA treatment. Articles were excluded if they were not available in English, not indexed in PubMed/MEDLINE or lacked sufficient methodological or clinical detail.

Study selection proceeded in two stages. First, titles and abstracts were screened for relevance. Second, full texts of potentially eligible articles were reviewed and assessed for inclusion in the final synthesis. Given the narrative design of the review, no formal PRISMA workflow, risk-of-bias scoring instrument, or meta-analysis was performed. Instead, the evidence was synthesized qualitatively, with emphasis on clinical relevance, conceptual coherence, study design, population characteristics, intervention characteristics, outcome measures, and major limitations.

This review was intended as an evidence-informed narrative synthesis rather than a fully systematic review. Accordingly, the goal was to provide a balanced and clinically meaningful overview for psychiatrists and addiction clinicians, while acknowledging the evolving and heterogeneous nature of the available literature.

3. Neurobiology of GLP-1 Use in Addictions

Glucagon-like peptide-1 (GLP-1) has emerged as a biologically plausible link between metabolic regulation and the neurocircuitry of addiction. Although GLP-1 receptor agonists (GLP-1RAs) were originally developed for type 2 diabetes and obesity, the reviewed literature indicates that GLP-1 signaling extends well beyond appetite and glycemic control, engaging central pathways involved in reward processing, cue reactivity, stress responses, relapse vulnerability, and executive control [5,6]. For psychiatrists, this is particularly relevant because it reframes substance use disorders (SUDs) not only as disorders of dopaminergic reinforcement, but also as conditions shaped by gut–brain communication, interoceptive signaling, and neurometabolic state [5,6].

GLP-1 is synthesized both in the periphery and in the central nervous system. Peripherally, it is secreted by enteroendocrine L-cells in response to nutrient ingestion; centrally, it is produced by preproglucagon neurons in the nucleus tractus solitarius (NTS) [5,6]. This dual origin is important because it provides two complementary routes through which GLP-1 signaling may influence addictive behavior. Peripheral GLP-1 can act through vagal afferents and circumventricular structures, whereas central GLP-1 neurons project directly to brain regions implicated in motivation, reward valuation, and behavioral control [5,6]. In this framework, GLP-1 is best understood as a neurometabolic signal capable of integrating bodily state with emotional and motivational behavior.

The anatomical distribution of GLP-1 receptors supports this interpretation. Across the reviewed articles, GLP-1 receptors are described in the hypothalamus, area postrema, NTS, ventral tegmental area (VTA), nucleus accumbens (NAc), prefrontal cortex (PFC), amygdala, and hippocampus [5,6]. These regions collectively subserve homeostatic control, salience attribution, reward learning, memory, affect regulation, and inhibitory control. Their involvement suggests that GLP-1R signaling may influence multiple clinically relevant dimensions of addiction, including craving, compulsive drug seeking, cue-induced relapse, stress-induced reinstatement, and impaired decision-making [5,6]. Amorim Moreira Alves et al. (2025) further highlight evidence of GLP-1 receptor expression in human frontal cortical regions, strengthening the translational relevance of these findings and raising the possibility that GLP-1-based interventions may affect higher-order processes such as impulse control, valuation, and executive regulation of behavior [5].

At the molecular level, GLP-1 receptors are class B G protein-coupled receptors that signal primarily through Gs-mediated activation of adenylate cyclase, increasing cyclic AMP and activating downstream pathways such as protein kinase A [5,6]. These intracellular events alter neuronal excitability, neurotransmitter release, and gene expression in circuits relevant to reinforcement and relapse [5,6]. Amorim Moreira Alves et al. (2025) also emphasize that GLP-1RA pharmacology is not uniform across compounds [5]. Different agonists may vary in central nervous system penetration, receptor internalization, and signaling bias between G-protein and β-arrestin-related pathways, which may influence both efficacy and tolerability [5]. This point is highly relevant to psychiatry, because compounds within the same therapeutic class may differ in their effects on mood, motivation, and reward sensitivity.

Within the mesocorticolimbic system, GLP-1 signaling appears to modulate reward in a nuanced rather than globally suppressive manner. In the VTA and NAc, GLP-1 receptor activation reduces the reinforcing properties of addictive substances and dampens drug-induced dopaminergic responses, but the reviewed literature suggests that this does not necessarily amount to indiscriminate dopamine blockade [5,6]. Instead, GLP-1RAs appear to recalibrate mesolimbic dopamine signaling in a context-dependent way, preferentially blunting drug-related or cue-driven hyperactivity while preserving basal reward function [5]. This distinction is clinically important, because treatments that broadly diminish reward processing may worsen anhedonia, apathy, and treatment adherence in psychiatric populations.

The synaptic mechanisms described in the reviewed articles are similarly complex. GLP-1 signaling modulates not only dopaminergic transmission but also glutamatergic and GABAergic activity within reward and relapse circuits [5,6]. In the VTA, GLP-1 receptor activation influences neuronal excitability in part through effects on glutamatergic input and local GABAergic tone [5]. In the NAc, GLP-1 signaling alters synaptic plasticity and reward-related responsiveness [5]. Marquez-Meneses et al. (2025) additionally highlight actions in the amygdala, hippocampus, and PFC, consistent with potential effects on emotional memory, stress-related responding, and top–down control over substance seeking [6]. Collectively, these findings support the view that GLP-1RAs may reduce craving and relapse not through a single neurotransmitter mechanism, but through coordinated modulation of the broader neural architecture of addiction [5,6].

A major conceptual contribution of the reviewed literature is the placement of GLP-1 within the gut–brain axis. Amorim Moreira Alves et al. (2025) describe vagal afferents as a principal route by which peripheral GLP-1 influences the brain, with vagal signaling to the NTS providing an entry point into central reward networks [5]. The NTS, in turn, projects to the VTA and NAc, establishing an anatomical pathway through which metabolic and visceral signals may shape motivation and reward seeking [5]. For psychiatrists, this is especially meaningful because craving is often experienced as a bodily, affectively charged state rather than a purely cognitive event. GLP-1 signaling may therefore act at the interface of interoception, autonomic tone, and motivational salience. Amorim Moreira Alves et al. (2025) further suggest that variation in vagal function may contribute to interindividual differences in treatment response, a point that may become increasingly relevant in future precision approaches to addiction treatment [5].

The manuscript by Marquez-Meneses et al. (2025) offers a particularly useful psychiatric formulation by applying the incentive-sensitization framework to GLP-1 biology [6]. According to this perspective, GLP-1RAs may influence both “liking,” understood as the hedonic pleasure associated with substance use, and “wanting,” understood as the motivational salience that drives compulsive seeking [6]. Preclinical studies show that GLP-1RAs reduce conditioned place preference as well as operant self-administration across several substances, suggesting effects on both hedonic reinforcement and motivated drug seeking [6]. Importantly, some of these findings indicate that GLP-1RAs can reduce drug-related “wanting” without producing overt aversion or suppressing all natural rewards [6]. This profile is attractive from a psychiatric standpoint because it raises the possibility of attenuating pathological craving without globally blunting motivation or pleasure.

The reviewed literature also suggests that GLP-1 signaling may act differently across phases of the addiction cycle. Marquez-Meneses et al. (2025) propose that during intoxication, GLP-1R activation primarily modulates mesolimbic circuitry, especially the VTA and NAc, thereby reducing the reinforcing and hedonic properties of substances [6]. During withdrawal, GLP-1RAs may exert effects through broader networks involving the lateral septum, hypothalamus, and brainstem, attenuating stress responsivity, negative affect, and somatic symptoms [6]. During anticipation and craving, modulation of prefrontal cortical regions and downstream targets such as the hippocampus and NAc may decrease cue-induced reinstatement and relapse vulnerability [6]. This phase-based model is particularly useful for psychiatric practice because it maps GLP-1 neurobiology onto symptom clusters that clinicians routinely encounter: reinforcement during active use, dysphoria and stress during withdrawal, and cue-driven compulsive ideation during remission [6].

Among specific substances, the strongest and most consistent evidence concerns alcohol. Both articles 5 and 6 describe robust preclinical findings showing that GLP-1RAs reduce alcohol intake, attenuate alcohol-induced dopamine release in the NAc, suppress conditioned place preference, reduce locomotor sensitization, and diminish relapse-like behavior [5,6]. These effects involve multiple nodes of the reward and relapse network, including the VTA, NAc, NTS, and other connected structures [5,6]. Early human studies cited in these reviews provide preliminary support for clinical relevance, with signals suggesting reductions in alcohol craving, alcohol self-administration, or alcohol-related neural reactivity, particularly in individuals with obesity or metabolic dysregulation [5,6]. For psychiatrists, this is notable because it suggests that GLP-1-based treatments may be especially promising in patients with co-occurring addictive and metabolic pathology.

Nicotine-related findings also support a neurobiological role for GLP-1 in addiction, although the circuitry may extend beyond the canonical mesolimbic pathway. Marquez-Meneses et al. (2025) highlight the NTS–medial habenula–interpeduncular nucleus axis as an important pathway in nicotine models, suggesting that GLP-1 signaling may influence both reward- and aversion-related processing in tobacco use disorder [6]. GLP-1RAs reduce nicotine intake, nicotine-seeking behavior, and some forms of reinstatement in preclinical studies, and they may also mitigate withdrawal-associated hyperphagia [6]. These findings are clinically relevant because weight gain and metabolic dysregulation are common barriers to smoking cessation and may be especially prominent in psychiatric populations.

For psychostimulants such as cocaine and amphetamine, the reviewed preclinical evidence indicates that GLP-1RAs reduce drug-induced locomotor activation, conditioned place preference, and other reward-related behaviors [5,6]. These effects are again associated with modulation of mesolimbic dopamine signaling and related circuitry [5,6]. However, the clinical evidence remains limited, and the translational pathway is less developed than in alcohol use disorder [6]. For opioids, Marquez-Meneses et al. (2025) describe a more heterogeneous literature, with some studies showing reduced opioid seeking, attenuation of withdrawal-related phenomena, and modulation of relapse-relevant circuits, while others report less consistent effects depending on the opioid or experimental paradigm [6]. Bade et al. (2025) add an important clinical perspective by linking GLP-1RA biology to chronic pain and opioid exposure, suggesting that these agents could eventually hold value in patient populations where pain, opioid use, and addiction risk overlap [7].

The article by Bade et al. (2025) is especially relevant for psychiatric readers because it broadens the significance of GLP-1 beyond classic addiction models [7]. It emphasizes that GLP-1RAs may influence inflammation, oxidative stress, macrophage polarization, microglial activity, and dopaminergic signaling, thereby affecting both chronic pain states and substance-related behavior [7]. This is important because chronic pain and SUDs frequently coexist, particularly in patients exposed to long-term opioid treatment. From a psychiatric perspective, agents capable of simultaneously reducing pain burden, lowering opioid dependence risk, and modulating reward circuitry would be of considerable interest [7]. Although the article focuses primarily on pain management, its discussion of GLP-1 effects on reward pathways and drug-seeking behavior supports the broader view that GLP-1RAs may be clinically useful in complex patients with overlapping pain, addiction, and metabolic comorbidity [7].

The translational implications for psychiatry are promising but still preliminary. The reviewed evidence suggests that GLP-1RAs may be most effective in biologically enriched subgroups, particularly patients with obesity, insulin resistance, or other metabolic disturbances [5,6]. This observation may reflect a true neurometabolic interaction rather than a simple consequence of weight loss. Indeed, Bade et al. (2025) explicitly note that beneficial effects in pain and related behavioral domains may occur independently of major weight change [7]. If confirmed, this would support the view that GLP-1RAs act directly on neural and inflammatory mechanisms relevant to addiction rather than merely improving metabolic health as a secondary route [7].

At the same time, the literature reviewed in articles 5 and 6 urges caution. The central effects that make GLP-1RAs attractive for addiction treatment may also create neuropsychiatric tolerability concerns [5]. Amorim Moreira Alves et al. (2025) note the possibility of adverse effects such as nausea, apathy, mood changes, anxiety, depression, and, rarely, suicidality, especially with compounds that have greater central penetration [5]. For psychiatrists, this point is essential. Any future use of GLP-1RAs in addiction treatment would require careful patient selection and systematic monitoring of mood, anxiety, motivation, eating behavior, and suicidal ideation [5]. The psychiatric safety profile may prove to be compound-specific, which further underscores the need to distinguish among GLP-1RAs rather than treating them as a homogeneous drug class [5].

In summary, the neurobiological evidence provides a plausible framework for understanding how GLP-1 receptor agonists may influence addiction-related processes, but the level of evidence differs across domains. Empirical data, particularly from preclinical studies, support the involvement of GLP-1 signaling in mesocorticolimbic reward pathways, gut–brain communication, cue reactivity, stress-related responses, and drug-seeking behavior. Human evidence, however, remains more limited and is currently strongest for alcohol-related outcomes, especially studies showing modulation of alcohol cue reactivity and reductions in some drinking-related measures. By contrast, broader claims regarding effects on executive control, interoception, relapse prevention, chronic pain–addiction overlap, or compound-specific psychiatric tolerability should be interpreted as mechanistic hypotheses rather than established clinical facts. Thus, the neurobiological section should be read as a translational rationale for further research, not as proof that GLP-1 receptor agonists are effective anti-addiction treatments in routine psychiatric practice [5,6,7].

4. Randomized Controlled Trials of GLP-1 Analogs in Addiction Treatment

Across the available randomized trials, GLP-1 analogs have been tested in two main clinical contexts: alcohol use disorder (AUD) and smoking cessation/nicotine dependence. The overall signal is more convincing for reduction in alcohol-related outcomes and for mitigation of post-cessation weight gain than for categorical smoking abstinence itself. Importantly, the trials differ substantially in population, co-interventions, duration, and endpoints, which likely explains part of the heterogeneity in the findings [8,9,10,11,12].

The first smoking-cessation signal came from a pilot randomized controlled trial of extended-release exenatide in 84 treatment-seeking smokers who were prediabetic and/or overweight. Participants were randomized to weekly exenatide 2 mg or placebo for 6 weeks; all also received nicotine patch treatment (21 mg) and brief smoking-cessation counseling. Exenatide increased 7-day point-prevalence abstinence at the end of treatment (46.3% vs. 26.8%; risk ratio 1.70, 95% credible interval 0.96–3.27), reduced end-of-treatment craving in the overall sample, reduced withdrawal symptoms among abstainers, and was associated with 5.6 pounds lower post-cessation body weight than placebo. Adverse events were infrequent, mild, and mainly limited to injection-site nodules, with no adverse event-related discontinuations. For psychiatrists, this study is notable because it suggested that a GLP-1 agonist might simultaneously target reinforcement-related processes and a clinically important relapse driver, namely weight gain after quitting smoking [12].

A larger and methodologically stronger smoking-cessation trial subsequently tested dulaglutide in 255 adult smokers with at least moderate nicotine dependence who wanted to quit. This was a 12-week, single-center, randomized, double-blind, placebo-controlled trial in which dulaglutide was added to standard care consisting of behavioral counseling and varenicline 2 mg/day. In contrast to the exenatide pilot, dulaglutide did not improve the primary smoking outcome: point-prevalence abstinence at week 12 was 63% with dulaglutide and 65% with placebo. Craving for smoking also declined substantially over time but did not differ between groups. However, dulaglutide clearly affected metabolic outcomes: weight decreased by 1.0 kg in the dulaglutide arm and increased by 1.9 kg in the placebo arm, yielding a baseline-adjusted between-group difference of −2.9 kg, and HbA1c also declined significantly with dulaglutide. Gastrointestinal symptoms were common, particularly nausea, but serious adverse events were judged unrelated to the study drug. Clinically, this trial suggests that in smokers already receiving an effective cessation package, dulaglutide may not add anti-craving or anti-relapse efficacy, yet it may still be valuable for limiting post-cessation metabolic burden [11].

A particularly informative extension of the dulaglutide smoking-cessation study was a predefined secondary analysis focused on alcohol use among participants who drank at baseline. This analysis included 151 participants from the parent trial who both consumed alcohol at baseline and completed 12 weeks of treatment. In this subgroup, dulaglutide was associated with a 29% lower alcohol intake at week 12 than placebo (relative effect 0.71, 95% CI 0.52–0.97), with the effect strengthening after additional adjustment for education. Notably, this reduction in drinking was not correlated with smoking status at week 12, suggesting that the alcohol signal was not simply a downstream consequence of successful smoking cessation. No effect was seen on other drug use, and the heavy-drinker subgroup was too small to show a clear between-group difference. For psychiatric interpretation, this is an important finding: within a trial that was negative for smoking abstinence, dulaglutide still appeared to reduce alcohol consumption, implying that GLP-1 signaling may affect alcohol-related reinforcement more robustly than nicotine abstinence outcomes in clinical populations. Because this was a secondary analysis of the same parent cohort, it should not be counted as independent evidence for smoking cessation efficacy, but it meaningfully strengthens the emerging AUD signal [8,11].

The first dedicated randomized clinical trial in treatment-seeking AUD tested once-weekly exenatide 2 mg over 26 weeks in 127 patients, all of whom also received standard cognitive-behavioral therapy. On the primary clinical endpoint, exenatide was not superior to placebo: both groups showed marked reductions in heavy drinking days and total alcohol intake, but between-group differences were not significant. However, the trial yielded some of the most compelling mechanistic data in this field. Exenatide significantly attenuated alcohol cue reactivity on fMRI, particularly in reward-related regions including the ventral striatum, and dopamine transporter availability was lower in the exenatide group than in the placebo group. In exploratory analyses, exenatide also reduced heavy drinking days and total alcohol intake in the subgroup with obesity (BMI > 30 kg/m²), whereas this effect was not seen in the overall sample. Adverse events were mainly gastrointestinal and injection site-related; importantly, no pancreatitis signal emerged. For psychiatrists, this trial is highly relevant because it suggests central target engagement even when the primary behavioral endpoint is negative, and it raises the possibility that the metabolic phenotype may modify responses. At the same time, the strong placebo response and the concurrent psychotherapy make efficacy interpretation more difficult [9].

The most encouraging alcohol trial to date in terms of overt behavioral outcomes is the phase 2 semaglutide study in 48 non-treatment-seeking adults with AUD. Participants received weekly semaglutide for 9 weeks using a low-dose escalation schedule (0.25 mg/week for 4 weeks, 0.5 mg/week for 4 weeks, and 1.0 mg for 1 week) or placebo. The primary endpoint was laboratory alcohol self-administration. Semaglutide significantly reduced the amount of alcohol consumed during the post-treatment self-administration session and lowered peak breath alcohol concentration, with medium-to-large effect sizes. Prospectively, semaglutide did not reduce average drinks per calendar day or the number of drinking days, but it did reduce drinks per drinking day, weekly alcohol craving, and heavy drinking over time relative to placebo. In a smoking subsample, semaglutide also predicted greater reductions in cigarettes per day. Body weight fell by about 5% in the semaglutide arm, and no serious adverse events, alcohol-related drug interactions, or treatment-related discontinuations were reported. From a psychiatric perspective, this trial stands out because it linked a clinically available GLP-1 agonist to reductions in both objective laboratory drinking and several clinically meaningful dimensional outcomes, even at relatively low doses [10].

Taken together, these trials suggest several patterns that are clinically and scientifically relevant. First, GLP-1 analogs appear more consistently associated with reductions in alcohol-related reinforcement, craving, or consumption than with improved smoking abstinence per se. Second, the most reproducible smoking-related benefit across studies is prevention of post-cessation weight gain, a clinically meaningful effect because weight gain is a common barrier to quitting and an important reason for relapse. Third, the psychiatric value of this drug class may lie at least partly in its effects on dimensional outcomes—cue reactivity, craving, drinks per drinking day, heavy drinking days, and weight-related relapse vulnerability—rather than only in binary abstinence outcomes. Fourth, metabolic phenotype may matter, although the evidence is not yet consistent: the exenatide AUD trial suggested benefits in obesity, whereas the semaglutide study did not replicate a clear BMI-dependent pattern [8,9,10,11,12].

Several limitations should temper clinical enthusiasm. Most trials were small or modest in size, several used active co-interventions that may have amplified placebo response or masked medication effects, and patient populations were selective. For example, the smoking studies were enriched for medical risk states such as overweight or prediabetes, whereas the semaglutide AUD trial excluded treatment-seeking participants and most non-alcohol substance use disorders; other studies excluded unstable psychiatric conditions. These design choices improve internal validity but limit direct generalization to typical psychiatric populations, where comorbidity, poly-substance use, and medication burden are common. Thus, the current evidence supports GLP-1 analogs as promising, mechanistically plausible candidates for addiction psychiatry, but not yet as established treatments for routine off-label use solely for addictive disorders [8,9,10,11,12].

At present, the strongest psychiatric rationale for further development lies in AUD. Exenatide provided neurobiological evidence of reduced alcohol cue salience despite a negative primary outcome, dulaglutide reduced alcohol intake in a smoking-cessation cohort, and semaglutide produced the clearest prospective reductions in alcohol self-administration, craving, and heavy drinking. In nicotine dependence, the clinical signal is mixed: exenatide showed promise in a pilot trial, but dulaglutide failed to improve abstinence when added to varenicline and counseling. Accordingly, the most defensible current position is that GLP-1 analogs are best viewed as emerging adjunctive treatments with potentially important transdiagnostic relevance for reward regulation, craving, and metabolic vulnerability, rather than as proven anti-addiction medications [8,9,10,11,12]. The summary of randomized controlled trials is presented in

Table 1. Summary of randomized controlled trials of GLP-1 analogs in addictions.

Author	Medication	N	Dose	Time of Intervention	Main Outcome	Main Limitation
[8]	Dulaglutide (predefined secondary analysis of smoking-cessation RCT)	151 baseline drinkers/completers (from parent cohort of 255)	Same parent trial regimen: 0.75 mg SC in week 1, then 1.5 mg SC weekly	12 weeks	Participants receiving dulaglutide drank 29% less alcohol than placebo at week 12; effect was not correlated with smoking status	Secondary analysis in smokers rather than a dedicated AUD trial; alcohol intake was self-reported and heavy-drinker subgroup was small
[9]	Exenatide once weekly	127 randomized	Exenatide 2 mg SC once weekly + standard cognitive-behavioral therapy	26 weeks	No significant reduction in heavy drinking days overall; however, reduced alcohol cue reactivity on fMRI and lowered dopamine transporter availability; exploratory benefit in obese subgroup	Primary clinical endpoint was negative, with substantial attrition and strong concurrent behavioral treatment/placebo response potentially obscuring efficacy
[10]	Semaglutide once weekly	48 randomized	0.25 mg/week for 4 weeks, 0.5 mg/week for 4 weeks, then 1.0 mg for 1 week	9 weeks	Reduced alcohol consumed during laboratory self-administration, drinks per drinking day, craving, and heavy drinking over time; no effect on average drinks/day or number of drinking days	Small, short phase 2 trial in non-treatment-seeking AUD participants, limiting generalizability
[11]	Dulaglutide	255 randomized	0.75 mg SC in week 1, then 1.5 mg SC weekly in weeks 2–12; all also received varenicline 2 mg/day + behavioral counseling	12 weeks	No improvement in smoking abstinence vs. placebo; significantly reduced post-cessation weight gain and HbA1c	Smoking trial with intensive concomitant treatment (varenicline + counseling), making a specific anti-addiction effect of dulaglutide harder to isolate
[12]	Exenatide adjunct to nicotine patch	84 randomized	Exenatide 2 mg SC once weekly + nicotine patch 21 mg daily	6 weeks	Improved 7-day point-prevalence smoking abstinence vs. placebo; reduced craving and withdrawal symptoms; lower post-cessation weight gain	Pilot trial in a selected sample of prediabetic and/or overweight smokers, limiting generalizability

.

5. Discussion

The contribution of this review should be understood as clinical rather than purely pharmacological novelty. While the neurobiological rationale and early clinical findings have been summarized elsewhere, psychiatrists require a cautious framework for applying this literature to real-world populations characterized by psychiatric comorbidity, poly-substance use, eating disorder vulnerability, suicidality risk, and complex medication burden. This review attempts to provide such a framework by emphasizing both the therapeutic promise and the psychiatric limits of the current evidence.

The present narrative review highlights a central tension in the current GLP-1 literature in addiction psychiatry: the neurobiological and translational rationale is strong, but strong clinical evidence remains too limited to justify firm conclusions regarding therapeutic efficacy in substance use disorders. This discrepancy is particularly evident when one compares the breadth of preclinical work with the relative scarcity of human trials. In the recent systematic review by Völker et al., a total of only five randomized controlled trials were included, and the available human data were concentrated mainly in alcohol and nicotine use disorders rather than across the broader spectrum of addictive conditions relevant to psychiatric practice [13].

Thus, at present, the field remains in an early translational phase. The currently available randomized studies suggest potentially meaningful effects on alcohol-related outcomes, craving, alcohol self-administration, and some weight-related dimensions relevant to smoking cessation, but the results are not yet sufficiently consistent to define GLP-1 analogs as established anti-addiction treatments. This interpretation is also supported by the broader review by Farokhnia et al. (2026), which concluded that randomized clinical trial data are still limited and mixed, even though preclinical work and observational evidence are encouraging [14]. In other words, the available literature supports promise, but not proof.

From a psychiatric point of view, this distinction is crucial. It would be premature to treat GLP-1 analogs as clinically validated treatments for addiction merely because they are biologically plausible and supported by early positive signals. The current evidence base remains constrained by modest sample sizes, short treatment periods, heterogeneous endpoints, selective populations, and frequent use of concurrent behavioral or pharmacological interventions. As a result, it is still difficult to determine whether the observed benefits reflect a direct and durable anti-craving effect, a broader modulation of reward-driven behavior, or a partial indirect effect mediated by metabolic change, weight loss, or altered interoceptive signaling. Völker et al. (2026) similarly emphasize that while preclinical findings are convincing, clinical results remain heterogeneous and limited by small samples and short study durations [13].

At the same time, caution is needed not only because efficacy remains uncertain, but also because psychiatric and behavioral safety cannot be regarded as a settled issue. The review by Strumila et al. (2024) is especially important in this regard [15]. Although it does not support a direct causal association between GLP-1 agonists and suicidal thoughts or behaviors, it raises a relevant clinical warning: the indirect consequences of rapid weight loss, biological stress, unmet expectations, and psychosocial readjustment may create vulnerability in susceptible individuals [15]. The authors argue that the observed psychiatric concerns may be better explained by confounding by indication and by the stress associated with metabolic and psychological change, rather than by a simple toxic effect of the drug itself [15].

This perspective is clinically useful for psychiatry. The question is not only whether GLP-1 analogs cause psychopathology in a direct pharmacological sense, but also whether they may destabilize vulnerable patients through secondary mechanisms. Abrupt appetite suppression, rapid shifts in eating behavior, changes in body image, disappointment in non-responders, or excessive psychological investment in weight loss may all be highly relevant in real-world psychiatric populations [15]. For patients with prior depression, disordered eating, obsessive-compulsive traits, trauma histories, or suicidality, these indirect pathways may matter more than formal pharmacovigilance signals. Thus, even in the absence of convincing evidence for direct psychiatric harm, psychiatrists should remain attentive to the broader psychological consequences of treatment [15].

An even more concrete warning emerges from the article by Richards et al. (2024), which frames GLP-1 treatment as potentially entering a clinical “danger zone” when appetite suppression becomes too effective and is not adequately monitored [16]. That article describes cases of severe restrictive intake, food aversion, dehydration, acute kidney injury, and clinically significant behavioral changes around eating, particularly in vulnerable patients and in inadequately supervised treatment contexts [16]. It further recommends proactive screening for disordered eating, close monitoring for maladaptive appetite-related responses, and ongoing assessment of broader health impacts during treatment [16].

These concerns are highly relevant for psychiatrists, because they broaden the frame beyond addiction outcomes alone. A medication that reduces craving or compulsive consumption may still become problematic if it also promotes restrictive eating, reinforces pathological control over intake, or triggers destabilizing changes in bodily experience and self-perception. This is particularly important in patients with eating disorder vulnerability, previous bariatric treatment, severe body image concerns, high reward sensitivity, or comorbid psychiatric illness [16]. In such patients, the line between therapeutic benefit and maladaptive overcorrection may become clinically thin.

At the same time, Farokhnia et al. (2026) appropriately reminds us that current evidence does not justify a purely alarmist interpretation [14]. That review notes that GLP-1 therapies do not presently appear to increase the risk of depression or suicidality overall, and that some recent studies even suggest beneficial mental health effects [14]. This is an important counterbalance. The psychiatric task, therefore, is not to reject GLP-1 analogs because of theoretical risk, but to place them within a framework of careful, individualized, and interdisciplinary care. The available evidence suggests neither clear psychiatric danger nor clear psychiatric validation; rather, it suggests a promising but still incompletely defined therapeutic space.

Where, then, should psychiatry go from here? First, larger and methodologically stronger randomized controlled trials are needed, particularly in clinically representative psychiatric populations with common comorbidities rather than narrowly selected research samples. Second, future studies should address not only efficacy but also dose, duration, discontinuation effects, tachyphylaxis, relapse after cessation, and predictors of response. Farokhnia et al. (2026) explicitly identify many of these unresolved issues, including duration of treatment, discontinuation, individual differences, mechanisms of action, psychiatric comorbidity, cost, and equitable access [14]. Third, psychiatrists should advocate for outcome measures that capture the complexity of treatment response, including craving, compulsive use, relapse vulnerability, eating behavior, mood, suicidality, hydration, nutritional status, and quality of life, rather than binary abstinence outcomes alone.

Until such data are available, GLP-1 analogs should be viewed in psychiatry as promising but unproven adjunctive tools. The strongest current signal appears to be in alcohol-related outcomes, whereas evidence for nicotine and other substance use disorders remains less definitive [13,14]. For now, their use should be guided by clinical caution, interdisciplinary collaboration, and structured monitoring rather than by therapeutic enthusiasm alone. Psychiatrists are well positioned to contribute to this next stage of the field: not by prematurely defining GLP-1 analogs as anti-addiction drugs, but by helping determine in whom, under what conditions, and at what psychiatric cost or benefit these agents may eventually find a place. A practical addiction psychiatry framework for GLP-1 receptor agonists should currently be conservative and monitoring-oriented rather than treatment-oriented. In patients who already have an established metabolic indication for GLP-1RA therapy, such as obesity or type 2 diabetes, psychiatrists may consider systematic observation of alcohol use, craving, smoking behavior, eating patterns, mood, suicidality, hydration, and weight trajectory during treatment. In contrast, in patients with substance use disorders but without a standard metabolic indication, GLP-1RAs should not yet be considered routine anti-addiction medications, because randomized controlled trial data remain limited, heterogeneous, and strongest only for selected alcohol-related outcomes. Particular caution is warranted in patients with eating disorder vulnerability, rapid weight loss, excessive appetite suppression, mood instability, or polysubstance use. Thus, the most practical current role of GLP-1RAs in addiction psychiatry is to generate clinically structured observations, support interdisciplinary collaboration, and identify subgroups for future trials, rather than to justify widespread off-label prescribing for substance use disorders.

Limitations

This review has several limitations. First, it was designed as a narrative review rather than a systematic review. Therefore, no formal PRISMA workflow, protocol registration, standardized risk-of-bias assessment, or meta-analysis was performed. The conclusions should be interpreted as a clinically oriented synthesis rather than a quantitative estimate of efficacy or safety.

Second, the literature search was restricted to English-language articles indexed in PubMed/MEDLINE, supplemented by manual screening of reference lists. As a result, relevant studies indexed in other databases, including Embase, Scopus, Web of Science, PsycINFO, Cochrane Library, regional databases, or non-English sources such as CNKI, may have been missed. This may introduce language and indexing bias.

Third, the available clinical evidence remains limited and heterogeneous. Randomized trials differ in sample size, duration, study population, GLP-1 receptor agonist used, dose, co-interventions, and outcome measures. These differences limit direct comparison between studies and reduce generalizability to routine psychiatric populations, where comorbidity, polysubstance use, eating disorder vulnerability, suicidality risk, and complex pharmacotherapy are common.

Finally, the field is evolving rapidly. New randomized trials, observational studies, and pharmacovigilance data may change the interpretation of the evidence soon after publication. Current data are still insufficient to define optimal patient selection, treatment duration, discontinuation strategy, long-term outcomes, or psychiatric and behavioral safety monitoring. Therefore, the conclusions of this review should support further research and careful clinical observation rather than routine off-label use of GLP-1 receptor agonists in addiction psychiatry.

6. Conclusions

For psychiatrists and addiction clinicians, GLP-1 receptor agonists should currently be viewed as promising but unproven adjunctive candidates rather than established treatments for substance use disorders. Their relevance to psychiatric practice lies not only in their possible effects on craving and reward processing, but also in the growing likelihood that patients receiving these medications for obesity or diabetes will report changes in alcohol use, smoking, appetite, compulsive behaviors, or mood during routine clinical care.

At present, the strongest clinical signal concerns alcohol-related outcomes, including reductions in alcohol cue reactivity, craving, alcohol self-administration, and some measures of heavy drinking. Evidence in nicotine dependence is less consistent and appears more clinically relevant for limiting post-cessation weight gain than for improving abstinence itself. For opioids, stimulants, and other substance use disorders, the evidence remains preliminary and is not sufficient to support routine clinical use.

In practical terms, psychiatrists should not interpret GLP-1 receptor agonists as stand-alone anti-addiction medications. Their potential use, if considered, should be embedded in standard addiction care, including psychosocial treatment, evidence-based pharmacotherapy where available, assessment of psychiatric comorbidity, and monitoring of substance use outcomes. Particular caution is needed in patients with eating disorder symptoms, rapid weight loss, excessive appetite suppression, dehydration risk, mood instability, suicidal ideation, or unrealistic expectations regarding treatment effects.

The immediate clinical value of this literature is therefore to inform careful observation, patient education, and interdisciplinary collaboration rather than premature widespread adoption. Future randomized trials should define which patients may benefit most, which substances and outcomes are most responsive, what doses and treatment durations are appropriate, and how psychiatric and behavioral safety should be monitored. Until such evidence is available, GLP-1 receptor agonists should remain investigational in addiction psychiatry, with their use guided by caution, structured monitoring, and collaboration between psychiatry, addiction medicine, endocrinology, and primary care.