Study links semaglutide and tirzepatide to fewer anxiety and depression diagnoses in obesity care

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Jul 1 2026

Real-world data suggest newer obesity drugs may carry different mental health profiles, with lower rates of several recorded psychiatric diagnoses but a cautionary anxiety and insomnia signal for tirzepatide versus semaglutide in people without diabetes.

Study: Neuropsychiatric association of tirzepatide and semaglutide in obesity with and without type 2 diabetes. Image Credit: Love Employee / Shutterstock

In a recent study published in the journal Communications Medicine, a group of researchers examined associations between initiation of tirzepatide or semaglutide and incident neuropsychiatric diagnoses in adults with obesity, compared with other weight-loss treatments, with separate analyses for individuals with and without type 2 diabetes.

Background

Currently, obesity affects more than 650 million adults globally, with implications for physical and mental well-being. People with obesity are more likely to suffer from psychiatric conditions, including anxiety and depression, than people without obesity. Moreover, mental illness can increase the chances of developing obesity. Newer incretin-based medications like semaglutide and tirzepatide have changed obesity treatment by producing substantial weight loss, but their effects on neuropsychiatric health remain uncertain in routine clinical practice. 

About the study

Researchers conducted a retrospective cohort study using de-identified electronic health records from the TriNetX United States Collaborative Network between January 2020 and November 2025. Adults meeting the study’s obesity criteria, who newly initiated tirzepatide or semaglutide, were identified and compared with matched individuals who started naltrexone-bupropion, phentermine, or phentermine-topiramate. The study included two separate analyses based on whether or not the participants had type 2 diabetes, and bariatric surgery recipients and a non-pharmacological weight-loss cohort were also evaluated as contextual comparison groups.

To increase comparability between treatment groups, propensity score matching was performed using demographic features, body mass index (BMI), glycated hemoglobin (HbA1c), medical history, co-existing treatments, healthcare utilization, and results of laboratory investigations. Patients with type 1 diabetes, prior bariatric surgery, history of organ transplantation, human immunodeficiency virus (HIV) infection, and end-stage renal disease were not included. Patients were monitored for a maximum of 24 months from treatment initiation until the initial diagnosis of a neuropsychiatric outcome, death, loss to follow-up, or end of study.

The primary outcome was a new diagnosis of anxiety disorder. Secondary outcomes included diagnoses of depression, mood disorder, insomnia, broadly defined cognitive deficits, nerve disorders, peripheral neuropathy, and substance-related disorders. Hazard ratios with 95% confidence intervals (CIs) were obtained using Cox proportional hazards models, adjusted for possible confounders, and various sensitivity analyses were also performed.

Study results

Before matching, individuals initiating incretin-based therapies differed considerably from comparator groups, particularly regarding psychiatric illnesses and psychotropic medication use. After matching on propensity scores, the groups had very similar baseline demographic characteristics, cardiometabolic health issues, laboratory measures, and health care use, allowing for more reliable comparisons between the treatment groups.

Individuals receiving semaglutide and tirzepatide generally had lower hazards of several recorded neuropsychiatric diagnoses when compared to bariatric surgery recipients.

Among individuals with obesity and type 2 diabetes, both medications were associated with lower hazards of anxiety disorders, broadly defined cognitive deficits, mood disorders, and depression. Similar results were seen for individuals without type 2 diabetes, but the neurological outcomes were less consistent between the groups. 

Semaglutide was associated with a higher hazard of nerve disorders in participants without diabetes, while tirzepatide showed lower hazards of substance-related disorders and peripheral neuropathy compared with bariatric surgery. The negative control analyses did not find any clinically meaningful differences, providing some reassurance about the analytical approach.

Compared with naltrexone-bupropion, semaglutide was associated with lower hazards of anxiety disorders, depression, cognitive deficits, and insomnia in adults with type 2 diabetes. Tirzepatide also showed lower hazards of anxiety disorders, depression, mood disorders, and cognitive deficits within this group.

Similar reductions were observed among adults without diabetes, indicating that both medications were consistently associated with lower rates of several recorded psychiatric diagnoses than this comparator.

Among adults without type 2 diabetes, a direct comparison between tirzepatide and semaglutide showed that tirzepatide was associated with higher hazards of anxiety disorders and insomnia than semaglutide, while no consistent differences were observed for most neurological outcomes after statistical adjustment. However, the authors noted that absolute and relative estimates differed for anxiety and insomnia in this head-to-head comparison, so these small differences should be interpreted cautiously.

Absolute risk differences indicated larger reductions in anxiety disorders among individuals with type 2 diabetes than among those without diabetes when tirzepatide or semaglutide was compared with naltrexone-bupropion.

Sensitivity analyses generally supported the primary findings. Per-protocol and as-treated analyses showed similar directions of association for anxiety and depression, although the strength of some associations varied depending on treatment persistence. Lower use of psychiatric medications, particularly selective serotonin reuptake inhibitors (SSRIs), was also observed among incretin-based therapy recipients compared with matched comparator groups.

Subgroup analyses suggested that the associations were generally consistent across age, BMI, and clinical characteristics, with some stronger associations observed among older adults and individuals with severe obesity.

Exploratory analyses generally showed greater reductions in body weight and HbA1c over 24 months in the incretin-based therapy groups than in the conventional anti-obesity medication groups, although some comparisons varied by diabetes status and outcome measure.

Additional sensitivity analyses using stricter outcome definitions and alternative cohort criteria yielded findings that were directionally similar, though some estimates became less precise due to smaller sample sizes. Importantly, cognitive findings were sensitive to outcome definition, and the broad cognitive endpoint included codes that could reflect acute or transient altered mental status rather than dementia or progressive cognitive decline.

Conclusion

The findings suggest that initiation of tirzepatide and semaglutide is associated with different neuropsychiatric outcome patterns depending on the comparator treatment and the presence of type 2 diabetes. Compared with naltrexone-bupropion, both medications were generally associated with lower hazards of anxiety and depression. In contrast, tirzepatide showed higher hazards of anxiety and insomnia than semaglutide among adults without type 2 diabetes.

These findings provide clinicians and patients with important real-world evidence to inform treatment decisions, while underscoring the need for continued patient monitoring. However, the results should not be interpreted as evidence that these medications prevent dementia, progressive cognitive decline, or psychiatric disease, because the study relied on diagnosis codes and remains vulnerable to residual confounding and misclassification.

In addition, as this study was observational, prospective studies with standardized neuropsychiatric assessments are needed to confirm these associations. Further research is needed to understand how these medications may influence mental health outcomes across different patient populations.

Want to read later? Download your PDF copy by clicking here.