Background: The Intersection of Diabetes and Mental Health
The comorbidity between type 2 diabetes and mental health disorders represents one of the most pressing challenges in modern medicine. Epidemiological data consistently demonstrate that individuals with diabetes face a substantially elevated risk of developing depression, anxiety, and even suicide. This bidirectional relationship creates a complex therapeutic landscape where managing blood glucose must be balanced against mental health stability.
GLP-1 receptor agonists have emerged as cornerstone medications for both diabetes and obesity management. However, the psychiatric effects of these agents have remained controversial, with limited and sometimes contradictory evidence regarding their impact on mood, anxiety, and self-harm behaviors. The question of whether GLP-1 receptor agonists alleviate or potentially exacerbate mental illness has significant implications for clinical decision-making in this vulnerable population.
This national cohort study from Sweden addresses this critical knowledge gap by examining the association between GLP-1 receptor agonist use and worsening mental illness in patients already diagnosed with depression, anxiety, or both. The research provides compelling evidence that certain GLP-1 receptor agonists, particularly semaglutide, may offer unexpected benefits for mental health outcomes.
Study Design and Methodology
Researchers conducted a nationwide cohort study using Swedish electronic health registers, identifying individuals with documented depression or anxiety disorder who initiated any antidiabetic medication between 2009 and 2022. The study employed a within-individual design, comparing periods of medication use versus non-use within the same individual, which effectively controls for time-invariant confounding factors such as genetics, socioeconomic status, and baseline health behaviors.
The analytical approach compared GLP-1 receptor agonists both as a group and individually against non-use of these agents, as well as against other second-line antidiabetic medications. This dual comparison strategy provides a comprehensive assessment of relative psychiatric safety and potential benefits.
The primary outcome was a composite measure of mental illness worsening, defined as psychiatric hospitalization, sick leave exceeding 14 days for psychiatric reasons, hospitalization due to self-harm, or suicide death. Secondary outcomes included worsening of depression and anxiety analyzed separately, as well as substance use disorder exacerbation and self-harm events.
Within-individual stratified Cox proportional hazards models generated adjusted hazard ratios with 95% confidence intervals, ensuring robust statistical inference while maintaining the study’s observational nature. Notably, a person with lived experience of these conditions contributed to both study design and manuscript preparation, enhancing the research’s patient-centered perspective.
Key Findings: Differential Effects Across GLP-1 Receptor Agonists
The cohort included 95,490 individuals, with 56,976 (59.7%) female and 38,514 (40.3%) male participants. The mean age was 50.6 years with a standard deviation of 12.3 years. Among these participants, 22,480 individuals received GLP-1 receptor agonists during the follow-up period.
Compared with non-use of GLP-1 receptor agonists, striking differences emerged across individual medications:
Semaglutide demonstrated the most robust association with reduced mental illness worsening, showing a hazard ratio of 0.58 (95% CI 0.51-0.65), representing a 42% reduction in risk. Liraglutide also showed a significant protective association, though more modest, with a hazard ratio of 0.82 (95% CI 0.76-0.89). In contrast, exenatide (HR 1.01, 95% CI 0.69-1.46) and dulaglutide (HR 1.01, 95% CI 0.85-1.20) showed no significant association with mental illness outcomes.
Semaglutide’s benefits extended across multiple psychiatric domains. The medication was associated with decreased risk of worsening depression (HR 0.56, 95% CI 0.44-0.71), reduced anxiety worsening (HR 0.62, 95% CI 0.52-0.73), and notably, lower risk of substance use disorder exacerbation (HR 0.53, 95% CI 0.35-0.80). Liraglutide demonstrated protective effects only for depression worsening (HR 0.74, 95% CI 0.64-0.87).
When examining GLP-1 receptor agonists as a therapeutic class, the group showed a significant association with reduced self-harm events (HR 0.56, 95% CI 0.34-0.92), providing additional evidence for mental health benefits.
Expert Commentary: Interpreting the Findings
These results carry significant implications for clinical practice, particularly for the substantial proportion of patients with diabetes or obesity who also struggle with depression and anxiety. The differential effects observed between semaglutide and other GLP-1 receptor agonists suggest that not all medications in this class should be considered equivalent regarding psychiatric outcomes.
Several mechanistic hypotheses may explain semaglutide’s apparent mental health benefits. Beyond its established effects on glucagon-like peptide-1 receptors in pancreatic beta cells, semaglutide may exert direct or indirect effects on brain regions involved in mood regulation, appetite suppression, and reward processing. The medication’s superior efficacy in weight reduction compared to other GLP-1 agonists might contribute indirectly to improved mental health through enhanced body image, reduced metabolic stigma, and improved physical functioning.
The finding regarding reduced substance use disorder worsening is particularly noteworthy, given the high comorbidity between substance abuse and mental health conditions. This observation warrants further investigation in dedicated studies examining semaglutide’s potential role in addiction medicine.
The study’s strengths include its large sample size, national scope, use of within-individual comparisons to minimize confounding, and comprehensive outcome definitions spanning hospitalization, work disability, and mortality. However, limitations should be acknowledged. Ethnicity data were unavailable, precluding assessment of potential disparities. The observational design cannot establish causality, and unmeasured time-varying confounders may still influence results. Additionally, the Swedish healthcare context may limit generalizability to other national settings with different healthcare systems or population characteristics.
Conclusion: A New Therapeutic Paradigm Emerges
This landmark Swedish cohort study provides the most comprehensive evidence to date that semaglutide and, to a lesser extent, liraglutide are associated with reduced risk of mental illness worsening in patients with depression and anxiety. For clinicians managing the care of individuals with co-occurring diabetes, obesity, and mental health conditions, these findings suggest that semaglutide may represent a particularly advantageous therapeutic choice, potentially offering dual benefits for both metabolic and psychiatric health.
The differential effects observed across GLP-1 receptor agonists highlight the importance of medication-specific rather than class-wide assumptions regarding psychiatric safety and efficacy. Healthcare providers should consider mental health outcomes alongside glycemic control and weight management when selecting among available therapeutic options.
The investigators appropriately call for randomized controlled trials to confirm these observational findings. Such trials would provide level I evidence regarding semaglutide’s psychiatric effects and could potentially establish a new therapeutic indication for this widely-used medication class.
Funding for this research was provided by the Sigrid Jusélius Foundation, Jane and Aatos Erkko Foundation, and the Finnish Ministry of Social Affairs and Health.
References
Taipale H, Taylor M, Lähteenvuo M, Mittendorfer-Rutz E, Tanskanen A, Tiihonen J. Association between GLP-1 receptor agonist use and worsening mental illness in people with depression and anxiety in Sweden: a national cohort study. The Lancet. Psychiatry. 2026 Apr;13(4):327-335. PMID: 41862258.