Glucagon-like peptide-1 receptor agonists and mental health: a narrative review of emerging benefits and risks
Article information
Abstract
Glucagon-like peptide-1 receptor agonists (GLP-1 RAs), originally developed for type 2 diabetes mellitus and obesity, are increasingly recognized for their significant impact on the central nervous system, leading to reports of both beneficial and adverse mental health effects. This review summarizes the current evidence on the effects of GLP-1 RAs on various psychiatric and neurocognitive conditions to evaluate their clinical benefits and potential risks. The literature has revealed a complex and multifaceted psychiatric profile. For depression and anxiety, the evidence is conflicting, with large observational studies showing contradictory results that are largely attributable to confounding by indication or methodological differences in the study design. In contrast, consistent and positive evidence suggests therapeutic potential for substance use disorders, particularly alcohol use disorders. Furthermore, emerging data indicate a significant neuroprotective role, with several cohort studies indicating a reduced risk of dementia. The major public and regulatory attention regarding suicidality appears to be driven by the methodological limitations of the initial reports, as well-controlled active comparator studies have not found an increased risk. However, the safety of GLP-1 RAs in high-risk psychiatric populations has not been established. In conclusion, while GLP-1 RAs show considerable therapeutic potential, their unresolved safety profile in patients with preexisting psychiatric conditions necessitates a cautious clinical approach. Future large-scale randomized controlled trials that include psychiatric populations are crucial for clarifying the direct neuromodulatory effects of these agents and establishing guidelines for their safe use.
Introduction
Glucagon-like peptide-1 receptor agonists (GLP-1 RAs), which mimic the action of the incretin hormone GLP-1, are a class of medications developed for the treatment of type 2 diabetes mellitus (T2DM) and obesity. The physiological functions of GLP-1 include stimulating insulin secretion from the pancreas, suppressing glucagon secretion, delaying gastric emptying, and acting on the brain to increase satiety. Through these multifaceted mechanisms, GLP-1 RAs have demonstrated robust efficacy in glycemic control and weight loss. The GLP-1 RAs approved by the U.S. Food and Drug Administration (FDA) and widely used include semaglutide, liraglutide, dulaglutide, and exenatide [1,2].
With the rapid increase in GLP-1 RA use, studies on effects other than weight management are emerging. Clinical studies have revealed that GLP-1 RAs provide significant benefits in extra-pancreatic areas, such as reducing cardiovascular disease risk, delaying the progression of diabetic kidney disease, and treating non-alcoholic fatty liver disease [3,4]. Furthermore, a growing body of evidence suggests that these drugs significantly affect the mental health of users. Many individuals have reported that various psychiatric symptoms, such as depression, anxiety, and addictive behaviors, either improve or worsen after using GLP-1 RAs [5]. Some studies have indicated that depression and anxiety scores are lower in patients with T2DM treated with GLP-1 RAs than in those treated with insulin [6]; however, concerns have been raised about serious psychiatric side effects, such as depression and suicidal ideation [7,8]. These conflicting results can cause confusion for both clinicians and patients. Therefore, this review aimed to comprehensively analyze the reported effects of GLP-1 RAs on mental health and evaluate their clinical benefits and safety.
Neurobiological background
GLP-1 receptors are widely distributed throughout the central nervous system, and GLP-1 is endogenously produced in areas such as the brainstem and hypothalamus. This provides a strong biological basis for the potential influence of GLP-1 RAs on mental functions. These receptors are primarily concentrated in two key circuits. First, they are found in homeostatic and reward centers, including the hypothalamus and ventral tegmental area, suggesting that they regulate not only simple appetite suppression, but also the motivational and hedonic aspects of food intake. Second, their distribution in cognitive and emotional circuits, such as the hippocampus and amygdala, supports their potential involvement in memory and mood regulation [1,9,10] (Fig. 1).
Schematic of neural circuits modulated by glucagon-like peptide-1 receptor agonists. The reward circuit, a neural pathway extending from the ventral tegmental area (VTA) to the nucleus accumbens (NA), modulates motivation and reward-seeking behavior. The VTA also projects to the prefrontal cortex (PFC), which works in conjunction with the amygdala (Am) and hippocampus (Hi) to perform functions related to memory and emotion. Hy, hypothalamus; NTS, nucleus tractus solitarius. Created with BioRender.com by the author.
Furthermore, GLP-1 RAs exert direct neuroprotective and regulatory effects at the cellular level. The key mechanisms include anti-inflammatory effects that reduce neuroinflammation, antiapoptotic effects that promote neuronal survival, and stabilization of mitochondrial function and reduction in oxidative stress, which improve energy metabolism [11,12]. Additionally, their ability to promote neurogenesis and enhance synaptic plasticity in the hippocampus, the center of learning and memory, and modulate the effects of major neurotransmitters, such as dopamine and serotonin, provides a neurobiological foundation for their effects on mood, motivation, and cognition [4,10,12,13].
Effects on depression and anxiety
The effects of GLP-1 RAs on depression and anxiety disorders are among the most actively debated topics. Conflicting results from large-scale studies have caused considerable confusion in clinical practice.
1. Conflicting findings on the risk of depression and anxiety
Some studies suggest that GLP-1 RAs can improve the symptoms of depression and anxiety. One study indicated that patients with T2DM treated with GLP-1 analogs had lower depression and anxiety scores than those treated with insulin [6]. Supporting this observation, a large-scale database analysis by Epic Research indicated that patients prescribed GLP-1 RAs were significantly less likely to be diagnosed with depression and anxiety disorders than those in the non-user group. However, the scientific validity of Epic Research reports is limited, as they are not peer-reviewed, and their methodology is unclear. Therefore, these findings should be regarded as supportive rather than confirmatory evidence [14]. However, other studies have raised concerns that GLP-1 RAs may increase the risk of developing psychiatric disorders. A large-scale cohort study by Kornelius et al. [15] reported that in patients who were obese, treatment with GLP-1 RAs was associated with a significantly higher risk of developing major depressive disorder and anxiety disorders compared to the non-user group (Table 1).
Comparison of cohort studies on the relationship between GLP-1 RAs and symptoms of depression and anxiety
2. Potential reasons for conflicting findings
The reason these large-database studies reached opposite conclusions is likely due to subtle differences in study design and various confounding factors that are difficult to control.
1) Confounding by indication
Obesity and depression are very closely linked biologically, and it is widely known that each increases the risk of the other by approximately 50% [16]. If this underlying risk is not statistically adjusted, a spurious association between drug use and the occurrence of psychiatric problems may appear. This confounding effect has been clearly demonstrated in a study by Shapiro et al. [17]. In the crude analysis of this study, the use of GLP-1 RAs appeared to have more than double the risk of suicidality compared to the use of dipeptidyl peptidase-4 inhibitors (hazard ratio [HR], 2.08). However, after adjusting for confounding variables, such as underlying diseases and psychiatric history, the risk decreased and the significant association disappeared (HR, 1.02) (Table 1).
2) Choice of comparator group
Another factor that significantly influences study outcomes is the choice of the comparator group. Studies using an inactive comparator (patients not using the drug) were much more susceptible to confounding variables than studies using an active comparator (patients taking other types of obesity/diabetes mellitus [DM] medications). Studies with conflicting results have differed decisively in this regard.
3) Direct versus indirect effects
Currently, a major point of contention is the difficulty in clearly distinguishing whether the observed mood changes are due to a direct change in brain circuits through GLP-1 receptor modulation or an indirect psychological and physiological result of weight loss and improved physical health [18].
3. Meta-analysis results
A meta-analysis by Chen et al. [19], which combined six randomized controlled trials (RCTs), supported a positive effect, finding that GLP-1 RAs significantly reduced depression scale scores compared with the control group (standardized mean difference, –0.12). However, a critical point of discussion is whether this antidepressant effect represents a direct neuromodulatory action or an indirect consequence of improved metabolic health (e.g., weight loss, glycemic control). According to Chen et al. [19], a significant limitation of the existing literature is that none of the included RCTs performed a formal mediation analysis to disentangle these two pathways. Therefore, the extent to which the observed mood improvements are a secondary benefit of better physical health versus a primary effect on central nervous system circuits remains unclear. This ambiguity, combined with the fact that the number of included studies was small and none were designed with depression as a primary outcome, highlights a critical gap in the current evidence and underscores the need for future research specifically designed to investigate these mechanisms [19-25] (Table 1).
Effects on substance use disorders
In contrast to debates over depression and anxiety disorders, there is a relatively consistent accumulation of positive evidence regarding the effects of GLP-1 RAs on substance use disorders (SUDs).
1. Alcohol use disorder
Alcohol use disorder (AUD) is currently one of the most actively researched areas regarding the therapeutic potential of GLP-1 RAs. Multiple large-scale observational studies have consistently indicated that GLP-1 RA use is significantly associated with a reduced risk of AUD incidence or recurrence. Specifically, studies by Qeadan et al. [26] and Wang et al. [27] found significant reductions in alcohol intoxication and in new-onset and recurrence risks of AUD (Table 2). Notably, a “within-individual design” study by Lähteenvuo et al. [28], which compared periods when the same patient was taking the drug versus not taking it, showed that the risk of alcohol-related hospitalization was significantly reduced during the medication period, suggesting a potent therapeutic effect in clinical settings (Table 2). Although the results from RCTs are mixed and require further research [29-31], consistent reports of reduced alcohol-related risks across various study designs and populations are promising.
Summary of studies on glucagon-like peptide-1 receptor agonists for substance use disorder
2. Nicotine use disorder (smoking)
The effects on smoking have been evaluated from two perspectives. While RCT results on the success of smoking cessation itself have been conflicting, the most important finding consistently observed across several studies is that GLP-1 RAs significantly suppress “post-cessation weight gain,” one of the major barriers to quitting. This is a unique advantage that no other smoking cessation treatment possesses, potentially making it a very effective treatment option for patients who hesitate to quit due to concerns about weight gain [32-35].
3. Other addictive substances
Research on other addictive substances is also actively underway (Table 2).
1) Opioids
Opioids are a class of drugs that act on opioid receptors in the brain and are primarily used for pain relief. This class includes prescription medications such as oxycodone and hydrocodone and illicit substances such as heroin and fentanyl. They have a high potential for addiction due to their ability to produce profound euphoria and analgesia. Although preclinical results are inconsistent [36], a large-scale analysis of patient records found that patients prescribed GLP-1 RAs had a significantly lower risk of future opioid overdoses [26].
2) Cannabis
Cannabis, derived from cannabis plants, is the most used psychoactive substance worldwide. Its primary psychoactive component, delta-9-tetrahydrocannabinol, causes euphoria and relaxation, and alters sensory perception. Although cannabis is often used recreationally and medicinally, its use can lead to the development of cannabis use disorder. A retrospective cohort study indicated that semaglutide use was associated with a reduced risk of cannabis use disorder incidence and recurrence compared with other obesity or DM treatments [37].
3) Cocaine
Cocaine is a potent central nervous system stimulant derived from the leaves of the coca plant. It primarily functions by blocking the reuptake of dopamine, norepinephrine, and serotonin, leading to intense euphoria, energy, and alertness. Owing to its powerful effects on the brain's reward pathway, it has a high potential to cause addiction. A small-scale RCT showed no significant changes in cocaine use or subjective effects, indicating the need for further research [38].
Neurocognitive effects
GLP-1 RAs appear to have a significant impact not only on mood and motivation, but also on cognitive function. Their potential in the prevention and treatment of neurodegenerative diseases has attracted considerable attention.
1. Dementia risk reduction
One of the most anticipated effects of GLP-1 RAs is their potential to prevent dementia. Several large-scale cohort studies have consistently indicated that GLP-1 RA treatment significantly lowers the risk of developing major dementias, including Alzheimer disease, Lewy body dementia, and vascular dementia [39-42]. Semaglutide, in particular, showed a more pronounced risk-reducing effect than other classes of GLP-1 RAs, suggesting that it may have additional neuroprotective effects beyond those common to the drug class [41] (Table 3). These results suggest the potential of GLP-1 RAs as disease-modifying treatments to slow or halt the progression of major neurocognitive disorders for which there are currently no definitive treatments [11,43].
Summary of clinical studies on the neurological effects of GLP-1 RAs
2. Cognitive function enhancement
Research has also shown that GLP-1 RAs not only lower the risk of dementia but also enhance cognitive function. In a study of patients who were obese with pre-DM or early T2DM, liraglutide improved cognitive function scores, such as short-term memory [44], and brain imaging studies confirmed a significant increase in prefrontal cortex activity [45]. Notably, this improvement in cognitive function was not statistically correlated with changes in metabolic indicators such as weight or blood sugar levels. This strongly supports the possibility that cognitive enhancement from GLP-1 RAs is not an indirect effect of weight loss but a unique neurological effect acting directly on the brain [42,45]. However, the effect on overall cognitive function is not prominent in patients with established dementia, suggesting the benefit is most pronounced in patients with early-stage DM [46-48].
3. Parkinson disease treatment
The therapeutic potential of these compounds for Parkinson disease (PD) has been actively investigated. Some cohort studies have indicated that GLP-1 RAs are associated with a significantly lower incidence of PD than other oral antidiabetic drugs [39,49-51]. However, contrary to the improvements in motor symptoms observed in early clinical trials [52], a recently published large-scale phase 3 clinical trial showed no significant difference after 2 years of exenatide administration compared to placebo [53]. Because various studies have shown conflicting results, further research is required to determine the long-term effects on PD.
Other psychiatric applications
1. Impact on hedonic eating and binge behavior
GLP-1 RAs directly act on the brain's reward circuits that regulate not only homeostatic eating to replenish energy, but also hedonic eating driven by the pursuit of pleasure [36,54]. This mechanism raises the expectation that GLP-1 RAs could be new alternatives for treating binge eating disorder and bulimia nervosa. Systematic reviews and meta-analyses have indicated that GLP-1 RAs are promising agents to significantly reduce binge and emotional eating [55,56].
2. Potential for aggression modulation
In specific patient populations where dietary control is extremely difficult, such as those with autism spectrum disorder, GLP-1 RAs present new therapeutic possibilities. A case report described a patient with severe autism who, after treatment with liraglutide, showed a significant decrease not only in food-related obsessive thoughts and binging but also in aggression not directly related to food. Although this is a single case, it indicates that GLP-1 RAs may have a positive impact on the overall regulation of impulsivity and compulsive behaviors beyond controlling appetite [55,57].
3. Health management in patients with chronic mental illness
Antipsychotic drugs such as clozapine and olanzapine are known to cause significant weight gain and metabolic syndrome, which are major causes of morbidity and treatment non-adherence in patients with schizophrenia and bipolar disorder [58,59]. GLP-1 RAs may offer a new solution for these patients. Several RCTs in patients with schizophrenia who were obese and taking antipsychotics have shown that liraglutide or exenatide led to significant weight loss and improvements in metabolic indicators, such as fasting glucose and HbA1c levels, compared to placebo. Importantly, no significant worsening of psychiatric symptoms or other negative effects was observed during this process. This indicates that GLP-1 RAs could be an important tool to more effectively and safely manage the physical health of patients with chronic mental illness, thereby improving their quality of life [60-63].
Suicidality controversy
A global investigation of the link between GLP-1 RAs and suicidal ideation began in 2023, with a few cases reported in Iceland, prompting a safety review by the European Medicines Agency (EMA). This event and subsequent analyses of European pharmacovigilance data have significantly raised awareness of the psychiatric safety of GLP-1 RAs [7,64].
1. Regulatory agency responses and limitations of initial evidence
After reviewing reported cases, the FDA and EMA provisionally concluded that the available data were insufficient to establish a clear causal relationship between the use of GLP-1 RAs and suicidal thoughts or actions [65]. This is because of the inherent limitations of pharmacovigilance data, which are based on spontaneous adverse event reporting. Such data are highly susceptible to influences, such as reporting bias, and cannot be used to infer causality.
2. Evidence from randomized controlled trials and real-world evidence
1) Randomized controlled trials
Multiple meta-analyses of RCTs involving tens of thousands of patients have consistently shown no statistically significant differences in the incidence of suicide-related events between GLP-1 RA and placebo groups [56,66] (Table 4). However, these findings have the critical limitation of exclusion bias. A significant number of major clinical trials intentionally excluded patients with a history of psychiatric illnesses at the design stage. This means that while the results demonstrate safety in a healthy population with low psychiatric risk, they do not directly prove safety in the vulnerable populations we are most concerned about [56,66].
Methodological comparison of key evidence on GLP-1 RAs and suicidality risk
2) Real-world evidence
Large-scale observational studies including patients with psychiatric histories provide crucial information to address this controversy. The most important finding from these studies is that the results differ dramatically depending on the choice of the comparator group. For instance, a study by Kornelius et al. [15] that used an inactive comparator (untreated patients who were obese) suggested an increased risk of suicidal behavior in the GLP-1 RA group. In contrast, studies employing an active comparator design, which better mitigates confounding by indication, consistently found no increased risk. After sophisticated statistical adjustments, large-scale cohort studies by Shapiro et al. [17], Hurtado et al. [67], and Ueda et al. [68] concluded that the risk of suicidality was not higher in the GLP-1 RA groups than in the groups using other types of DM or obesity medications. Furthermore, Wang et al. [69] observed a protective effect of semaglutide and reported a significantly reduced risk of new-onset suicidal ideation (Table 4).
3. Core of the controversy and its implications
In conclusion, the “suicidality controversy” can be interpreted not as a clear drug side effect but as an issue arising from the conflict between different types of evidence and the inherent limitations of each research method. Therefore, the central issue of the controversy shifts from whether GLP-1 RAs directly cause suicidality to the more nuanced question of whether the available safety data can be generalized to high-risk psychiatric populations, particularly since these groups were often excluded from pivotal clinical trials. This debate highlights a broader challenge in the new drug development process, namely the need for more robust psychiatric safety assessments, particularly for vulnerable populations.
Conclusion
GLP-1 RAs exert complex effects in the psychiatric domain. Although conflicting results exist for depression and anxiety, making it difficult to draw clear conclusions, relatively consistent and positive results have been reported for SUDs, including AUD, and the potential for dementia prevention. GLP-1 RAs can also play a crucial role in improving treatment adherence and quality of life in patients with certain chronic mental illnesses by effectively managing the metabolic issues caused by antipsychotics. Regarding the most significant controversy regarding the risk of suicidality, most well-controlled studies have not observed an increase, but safety in high-risk groups with a history of psychiatric conditions has not yet been clearly established.
The following tasks are crucial for advancing this field. First, sophisticated research designs are essential to clearly distinguish whether the observed psychiatric effects are due to the drug’s direct neuromodulatory action or the indirect psychological consequences of weight loss. Second, large-scale, long-term RCTs that, unlike previous studies, actively include patients with a history of psychiatric illness and set psychiatric outcomes as primary endpoints are urgently needed. Third, head-to-head studies that directly compare the effects of different GLP-1 RAs such as semaglutide and tirzepatide on psychiatric symptoms are required. Finally, discovering genetic and biological biomarkers that can predict which patients will experience positive or negative psychiatric effects will be the cornerstone of personalized medicine.
Although GLP-1 RAs are highly effective in treating T2DM and obesity, their safety in patients with a history of psychiatric disorders has not been sufficiently studied. Therefore, when prescribing in a clinical setting, it is necessary to assess the patient's psychiatric risk factors (e.g., comorbid psychiatric disorders, history of self-harm, or suicide attempts). If a patient is deemed high-risk, it is safer to prescribe the medication cautiously only after fully explaining the potential risks, such as depression or suicidal ideation, to the patient or caregiver and obtaining consent for continuous monitoring.
Notes
Conflicts of interest
No potential conflict of interest relevant to this article was reported.
Funding
None.