Obesity is a prevalent and complex condition that significantly contributes to morbidity and mortality worldwide (1), with associated comorbidities including type 2 diabetes (T2D), cardiovascular disease, and a range of psychological disorders such as depression and anxiety (2–6). Despite advances in lifestyle and pharmacological interventions, effective long-term management of obesity remains challenging. Recently, tirzepatide, a novel dual glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1) receptor agonist, has demonstrated considerable efficacy in improving glycemic control and promoting weight reduction in patients with obesity (7–9). Its dual action mechanism offers a promising approach for the treatment of obesity and related metabolic conditions.

However, the psychological impact of tirzepatide, particularly its potential association with mental health outcomes, remains underexplored. This is of particular concern given the European Medicines Agency’s (EMA) recent alert regarding a possible link between GLP-1 receptor agonists and an increased risk of suicidal ideation and behavior (10, 11). Although randomized controlled trials (RCTs) have not definitively established a psychiatric risk profile for other GLP-1 receptor agonists such as liraglutide and semaglutide, observational studies have largely indicated no increased risk of suicide with their use (12–15). Nevertheless, only a few studies to date have specifically evaluated the risk of suicidal ideation or attempts associated with tirzepatide, leaving a significant gap in the literature that warrants further investigation (11). Additionally, it remains uncertain whether weight loss alone, regardless of the specific intervention, exerts a protective or detrimental effect on suicide risk, as prior data on this topic are conflicting (6, 16, 17).

In light of these concerns, this study aims to investigate the potential association between tirzepatide use and the risk of suicide in patients with obesity. By addressing this gap, we seek to contribute to a more comprehensive understanding of the safety profile of tirzepatide and inform clinical decision-making.

This study provides the first evidence supporting the safety of tirzepatide in relation to suicidal ideation and attempts. Over a median follow-up of 365 days, tirzepatide was associated with a 48% lower risk of suicidal ideation or attempts (aHR: 0.52, 95% CI: 0.28–0.91, p = 0.001). The consistent risk reduction observed at 3, 6, and 12 months further indicates a potential protective effect of tirzepatide over time.

Although the primary focus of our study was on patients with overweight or obesity, approximately 29% of the cohort also had T2D, a significant subgroup due to the well-established link between T2D, psychiatric disorders, and obesity. Our study demonstrated a consistently lower risk of suicide attempts in these patients. Subgroup analyses exploring the association between BMI, HbA1c, and GFR indicated that lower suicide ideation and attempt risks were observed across these metabolic variables. Despite the overall low incidence of suicide-related events, patients with better-controlled HbA1c levels and higher GFR appeared to have a reduced risk, highlighting the potential mental health benefits of improved glycemic control and kidney function in this population. Additionally, subgroup analyses (Figure 3) suggested directionally lower risks of suicidal ideation or attempts among female patients and those aged 45–64 years, whereas findings among individuals aged ≥65 years and those<45 years were neutral. These observations may be explained by limited statistical power due to very low event numbers, as well as competing mental health risk factors in older adults. Metabolic subgroups demonstrated similar patterns, with patients having HbA1c<7% or eGFR ≥60 mL/min/1.73 m² showing trends toward lower risk. These associations are biologically plausible given the links between metabolic health, quality of life, and psychological well-being; however, all subgroup findings should be regarded as exploratory.

Our findings align with previous observational studies on GLP-1 RAs, which generally report no significant association between these medications and increased suicide risk (12–15). Interestingly, Almeida et al. found that individuals exposed to GLP-1 RAs were at a greater risk of being prescribed antidepressants, suggesting a potential influence of GLP-1 RAs on mood. While this might seem contradictory to our findings, it is important to note that increased antidepressant use does not necessarily equate to increased suicidal ideation or attempts. In fact, antidepressant prescriptions may reflect proactive management of underlying depression, potentially reducing the risk of severe outcomes like suicide (21). Given the high prevalence of pre-existing depression and anxiety in our cohort, concomitant antidepressant or other psychotropic medication use is likely common. However, incomplete capture of medication adherence, dosage changes, and polypharmacy within TriNetX may introduce residual confounding.

The psychiatric safety observed in our study is consistent with findings from major RCTs involving tirzepatide, such as the SURMOUNT trials (7, 8, 22). These studies primarily focused on weight reduction and metabolic outcomes, but they also reported improvements in overall quality of life. While the primary outcomes of these trials did not specifically assess suicidal ideation or behaviors, the significant weight loss achieved with tirzepatide could have positively impacted patients’ psychological well-being, indirectly reducing the risk of mental health-related adverse outcomes. The substantial weight reduction observed in the SURMOUNT trials, up to 20%, is well-established to improve self-esteem, body image, and physical functioning, which are key factors in mitigating suicide risk. Our findings, combined with data from the SURMOUNT studies, suggest that tirzepatide’s comprehensive benefits on metabolic and psychological health make it a safe and effective option for patients with obesity, including those at risk for mental health issues.

The reduction in suicidal ideation and attempts observed in patients treated with tirzepatide could be explained by several biological mechanisms. GLP-1 receptor agonists, including tirzepatide, are thought to exert neuroprotective effects through the modulation of neuroinflammation, the hypothalamic-pituitary-adrenal (HPA) axis, and neurotransmitter pathways, which play key roles in mood regulation and stress response (23). Tirzepatide’s dual agonism, targeting both GLP-1 and GIP receptors, may provide additional benefits (24). While GLP-1 receptor agonists have shown potential in mood stabilization, the GIP pathway may further enhance emotional and cognitive outcomes, potentially contributing to the observed reduction in suicide risk (25). GIP signaling, though less studied in psychiatric contexts, has been linked to improved cognitive processes in animal models, and this dual pathway activation could offer unique neuropsychiatric benefits (26, 27).

The study has several strengths, including the use of a large, real-world database from the TriNetX network, which enhances the generalizability of our findings. We applied rigorous methodology, including PSM, and compared tirzepatide with other active anti-obesity medications to ensure balanced and robust comparisons. Moreover, the inclusion of a wide range of covariates, such as socioeconomic status and multiple comorbidities (e.g., depression, anxiety), enabled a comprehensive analysis. Finally, we evaluated suicide risk across multiple time frames, 3, 6, and 12 months, offering valuable insights into both the short- and intermediate-term effects of tirzepatide.

Nonetheless, our study has several limitations. First, as a retrospective observational analysis using EHR data, residual confounding is possible, even though we employed PSM and adjusted for known psychiatric and metabolic covariates, as well as socioeconomic status and lifestyle-related factors (smoking and alcohol use). However, we were unable to account for the severity of psychiatric illness or healthcare access, which may influence both treatment allocation and suicidality risk. Second, we excluded patients with prior suicidal ideation or suicide attempts to reduce confounding, but this prevented us from evaluating whether tirzepatide could reduce suicidality in that high-risk subset. Third, the low event rate, 17 cases in the tirzepatide group and 33 in the control group, limits statistical power and magnifies the impact of potential misclassification. In addition, reliance on diagnostic coding may underestimate the true incidence of suicidal ideation, as sensitivity and specificity of such codes are uncertain. Fourth, our follow-up period was only one year, which may not capture longer-term or delayed psychiatric events. In addition, we did not separately analyze censoring due to treatment discontinuation, switching, or death, and no competing-risks approach was applied. Given the low number of events and platform constraints, such analyses were not feasible. Fifth, while TriNetX spans multiple healthcare organizations, its data primarily derive from U.S. health systems and may not fully generalize to underserved or non-U.S. populations. Sixth, we did not perform imputation for missing data. Analyses were conducted using available values within the TriNetX platform, and patients with missing covariates were excluded from the relevant variable-specific analyses. This may introduce bias if missingness was not random. Seventh, consistent weight-loss data and patient-reported measures of mood or treatment satisfaction were not available in the TriNetX platform. This may limit our ability to evaluate whether the observed protective association is mediated by improvements in psychological well-being. Finally, our subgroup analyses did not use multiple-comparison corrections and should be regarded as exploratory, particularly given the small number of events. In addition, TriNetX’s privacy policy required rounding of event counts<10 to 10, which may slightly distort reported incidence rates but does not materially affect HR estimates derived from Cox models.

In conclusion, our results indicate that tirzepatide was not associated with an increased risk of suicidal ideation or suicide attempts. Although a HR below unity suggests a possible protective association, the absolute risk reduction was very small, and this finding should be considered tentative and hypothesis-generating. The key clinical message is that no safety signal for increased suicidality was observed. Longer-term prospective studies or randomized clinical trials with psychiatric endpoints are warranted to validate this association and to explore potential biological mechanisms in greater detail.