Tirzepatide and Semaglutide Show Similar Psychiatric Safety Over Two Years
A large real-world study finds tirzepatide and semaglutide carry comparable neuropsychiatric risk, while semaglutide outperforms older GLP-1 drugs.
Summary
Concerns about depression and suicidal thoughts linked to GLP-1 weight-loss drugs have grown as their use has exploded. This large retrospective study compared tirzepatide, semaglutide, and older GLP-1 receptor agonists in over 165,000 matched patient pairs drawn from a global database of 192 million records. Over two years, tirzepatide and semaglutide showed nearly identical rates of depression, anxiety, and suicidal ideation. Notably, semaglutide was associated with significantly lower psychiatric event rates than earlier GLP-1 drugs — including a 51% lower risk of suicidal ideation in year one. A marginally higher anxiety signal was seen with tirzepatide in year two, but researchers caution this may reflect statistical noise from multiple comparisons. Overall, the newest GLP-1 agents appear neuropsychiatrically safe relative to their predecessors.
Detailed Summary
Safety concerns around GLP-1 receptor agonists and mental health outcomes have intensified as these drugs move from niche diabetes treatments to mainstream obesity therapies used by tens of millions. Regulatory agencies have examined whether these medications increase risks of depression or suicidal thinking, making large real-world studies critical for clinical guidance.
This retrospective cohort study leveraged the TriNetX Global Federated Network, encompassing over 192 million patients. Adults with type 2 diabetes, obesity, or both who started tirzepatide, semaglutide, or older GLP-1 receptor agonists between July 2022 and June 2025 were enrolled under a new-user design with a 12-month washout period. Propensity score matching balanced baseline demographics, clinical variables, and metabolic factors, yielding 85,546 pairs for the tirzepatide-versus-semaglutide comparison and 80,115 pairs for semaglutide versus other GLP-1 agents. Cox proportional hazard models estimated risk across year one (days 31–365) and year two (days 366–730).
The headline finding is that tirzepatide and semaglutide were statistically indistinguishable on composite psychiatric outcomes at both time points. A nominally higher anxiety hazard with tirzepatide emerged in year two (HR 1.052), but investigators urge caution given the multiple comparisons performed. More striking was semaglutide's advantage over older GLP-1 drugs: 19% lower depression risk, 8.5% lower anxiety risk, and a remarkable 51% lower suicidal ideation risk in year one.
For clinicians and patients, these findings are reassuring. They suggest the neuropsychiatric safety concerns raised about earlier incretin therapies may be attenuated with newer agents, and that tirzepatide does not introduce additional psychiatric risk beyond semaglutide.
Key caveats include the observational design, potential residual confounding despite propensity matching, and reliance on diagnosis codes rather than validated psychiatric instruments. The summary is based on the abstract only, limiting full methodological appraisal.
Key Findings
- Tirzepatide and semaglutide had statistically identical composite psychiatric risk over 2 years.
- Semaglutide cut suicidal ideation risk by 51% versus older GLP-1 receptor agonists in year one.
- Semaglutide also reduced depression risk by 19% and anxiety by 8.5% vs. earlier GLP-1 drugs.
- A marginally higher anxiety signal with tirzepatide in year 2 requires cautious interpretation.
- Study covered 165,000+ matched patient pairs from a 192-million-patient global database.
Methodology
Retrospective cohort study using the TriNetX Global Federated Network (>192 million patients). New-user design with 12-month washout and propensity score matching on demographic, clinical, and metabolic variables. Cox models estimated hazard ratios across two annual landmark periods.
Study Limitations
Observational design limits causal inference and residual confounding cannot be excluded despite propensity matching. Psychiatric outcomes rely on diagnostic codes rather than validated clinical assessments. This summary is based on the abstract only, preventing full methodological review.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: