Semaglutide and Depression: Weight Loss vs Mental Health
GLP-1 and Mental Health
44%
In a Swedish cohort of 95,490 people with depression or anxiety, semaglutide use was associated with a 44% lower risk of worsening depression requiring hospital care or sick leave.
Taipale et al., The Lancet Psychiatry, 2026
Taipale et al., The Lancet Psychiatry, 2026
View as imageThe largest study ever conducted on GLP-1 drugs and mental health found that people taking semaglutide had a 44% lower risk of worsening depression. In the same year, a randomized clinical trial found semaglutide did not improve depressive symptoms at all. And individual case reports describe patients whose depression got measurably worse after starting the drug. All three of these findings can be true simultaneously, and understanding why is the point of this article. For the broader picture of how GLP-1 drugs affect psychological well-being, see our overview of body image and mental health after GLP-1 weight loss.
Key Takeaways
- A 2026 Swedish cohort study of 95,490 people found semaglutide was associated with 44% lower risk of worsening depression and 38% lower risk of worsening anxiety (Taipale et al., The Lancet Psychiatry, 2026)
- Post hoc analysis of the STEP 1, 2, 3, and 5 trials (n=3,377) found no increased risk of depression or suicidal ideation with semaglutide vs placebo over 68-104 weeks (Wadden et al., 2024)
- A 2026 RCT in 72 patients with major depressive disorder found semaglutide did not reduce depressive symptom severity over 16 weeks, though it improved cognition (Badulescu et al., 2026)
- A meta-analysis of GLP-1 agonist trials found no statistically significant increase in suicidal ideation or suicide completion (Alansari et al., 2025)
- A 12-month propensity-matched cohort of 25,000+ semaglutide users found reduced risk of depression, anxiety, and substance use disorders compared to matched controls (De et al., 2024)
- Individual case reports document worsened depression in specific patients, suggesting heterogeneous responses that population-level data may obscure (Manoharan et al., 2024)
The population-level evidence: large studies show benefit
Three large observational studies point in the same direction: across populations, semaglutide use is associated with lower rates of depression and psychiatric deterioration.
The most striking is the 2026 Swedish national cohort study published in The Lancet Psychiatry. Researchers analyzed data from 95,490 people with diagnosed depression or anxiety who used antidiabetic medications, with a mean follow-up of 5.2 years. During periods of semaglutide use, the risk of sickness absence or hospital care due to worsening psychiatric illness dropped 42% compared to periods when the same individuals were not using GLP-1 drugs. For depression specifically, the reduction was 44%. For anxiety disorders, 38%. For substance use disorders, 47%.
The within-person design is important. Rather than comparing semaglutide users to non-users (who might differ in dozens of ways), the study compared each person to themselves during periods on and off the medication. This eliminates many confounders that plague standard observational designs.
A 12-month propensity-score matched cohort study in eClinicalMedicine reached similar conclusions. De and colleagues (2024) matched over 25,000 semaglutide users with type 2 diabetes to non-users and found reduced risk of depression (HR 0.83, 95% CI 0.74-0.92), anxiety disorders, and substance use disorders at 12 months.[1]
Cooper et al. (2023) found that GLP-1 receptor agonist use was a protective factor against incident depression in patients with diabetes mellitus, with a hazard ratio suggesting meaningful clinical protection.[2]
These are not small effects. A 44% reduction in worsening depression is larger than what most antidepressant trials demonstrate.
The clinical trial evidence: safety confirmed, efficacy unclear
Randomized controlled trials tell a different story. Not a contradictory one, but a more nuanced one.
Wadden and colleagues (2024) conducted a post hoc analysis of psychiatric safety across the STEP 1, 2, 3, and 5 trials, which enrolled 3,377 participants without known major psychopathology for 68-104 weeks of treatment with semaglutide 2.4 mg weekly or placebo.[3] The findings were clear: semaglutide did not increase the risk of moderately severe depression symptoms (PHQ-9 score of 15 or higher) or suicidal thoughts/behavior compared to placebo. In fact, the semaglutide group showed modest improvements in depression scores.
But these trials excluded people with pre-existing major psychiatric conditions. They tell us semaglutide is safe for people without serious mental illness. They do not tell us whether it treats depression.
The first true test of that question came in 2026. Badulescu and colleagues randomized 72 adults with diagnosed major depressive disorder to receive 14 mg oral semaglutide or placebo for 16 weeks.[4] The primary outcome was change in depressive symptom severity.
Semaglutide did not improve depression scores. It did not increase suicidal ideation either. Where it did show effects was on cognition: secondary analyses found improvements in global cognitive function. Participants also lost an adjusted mean of 6.03 kg more than the placebo group.
This disconnect matters. The observational studies show that semaglutide users have less depression. The RCT shows semaglutide does not directly treat depression. The most likely reconciliation: the weight loss, improved metabolic function, and reduced "food noise" that semaglutide produces may indirectly improve quality of life and mood at a population level, even though the drug is not acting as an antidepressant per se.
For readers interested in how semaglutide alters the psychological experience of eating, see what "food noise" is and how GLP-1 drugs quiet it.
The suicidality question: what the data actually shows
The FDA investigated potential links between GLP-1 receptor agonists and suicidal thoughts in 2023, driven by case reports and pharmacovigilance signals. This investigation prompted a wave of research.
Alansari and colleagues (2025) published what they described as the first meta-analysis specifically assessing GLP-1 agonists and suicidal ideation.[5] Their conclusion: no statistically significant increase in suicidal ideation or suicide completion was associated with GLP-1 agonist use across the pooled trial data.
Wang et al. (2025) took a different approach, analyzing the FDA Adverse Event Reporting System (FAERS) database for depression and suicide/self-injury signals associated with semaglutide, liraglutide, and tirzepatide.[6] FAERS data is inherently limited (it captures reports, not confirmed causation), but it provides a pharmacovigilance lens that RCTs miss. Their analysis found signals worth monitoring but no clear evidence of a causal relationship.
Guirguis et al. (2024) analyzed FAERS data specifically for suicidal thoughts and self-injury in weight loss treatments, using metformin and orlistat as comparators.[7] Their findings similarly did not support a unique suicidality risk from GLP-1 agonists above background rates.
Tang et al. (2026) conducted a target trial emulation comparing GLP-1 receptor agonist use against SGLT2 inhibitors and DPP-4 inhibitors in adults with type 2 diabetes.[8] GLP-1 agonist initiation was not associated with increased risk of most psychiatric disorders, though a modest association with anxiety disorder was observed.
The weight of the evidence: GLP-1 agonists do not appear to increase suicidality risk at a population level. The FDA investigation findings align with this conclusion.
Individual cases of worsened depression
Population-level safety does not mean every individual responds the same way. Manoharan et al. (2024) published a case report in Innovations in Clinical Neuroscience describing a patient whose depression measurably worsened on Ozempic (semaglutide).[9]
This is not an isolated report. Zheng et al. (2025) integrated real-world pharmacovigilance data with Mendelian randomization analysis and found that while GLP-1 agonists as a class showed overall psychiatric safety, individual variation was substantial.[10] Some patients clearly deteriorate.
Several mechanisms could explain individual worsening:
Rapid weight loss and identity disruption. Losing 15-20% of body weight in a year changes how people relate to food, their bodies, and social situations. For some, this is liberating. For others, it triggers anxiety, disordered eating patterns, or depression related to loss of a coping mechanism. The psychological adjustment to GLP-1 weight loss is not uniformly positive.
Nausea and reduced eating. The gastrointestinal side effects of semaglutide, particularly in the dose-titration phase, can reduce caloric intake substantially. Caloric restriction alone can worsen mood in vulnerable individuals. For more on the GI experience, see nausea on semaglutide or tirzepatide.
GLP-1 receptors in the brain. GLP-1 receptors are expressed in multiple brain regions involved in mood regulation, including the hippocampus and amygdala. How semaglutide modulates these receptors likely varies by individual neurobiology.
Pre-existing vulnerability. The STEP trials excluded major psychopathology. People with treatment-resistant depression, bipolar disorder, or complex trauma histories were not studied in controlled settings. Their response to semaglutide is genuinely unknown.
The animal evidence: what preclinical models show
Preclinical research provides mechanistic clues that human studies cannot. De and colleagues (2024) demonstrated in a type 2 diabetes mouse model that semaglutide attenuated anxious and depressive-like behaviors and reversed cognitive impairment.[11] The proposed mechanisms included reduced neuroinflammation, improved insulin signaling in the brain, and modulation of serotonin and dopamine pathways.
These findings suggest that semaglutide's psychiatric effects are not purely mediated by weight loss. The drug appears to act directly on neural circuits involved in mood. But mouse behavior is not human depression, and these models cannot capture the psychological complexity of body image, social relationships, and identity that shape human mental health responses to weight loss drugs.
What the heart failure data adds
Balata et al. (2025) examined semaglutide's effects on mood specifically in obese patients with heart failure over one year.[12] This population is particularly relevant because heart failure and depression are strongly comorbid, and the metabolic burden of both conditions is severe.
The retrospective analysis found improvements in both health status and mood in semaglutide-treated patients. The significance of this finding is that it demonstrates mood benefits in a medically complex population, not just the relatively healthy participants enrolled in obesity trials.
Why the evidence seems contradictory (but is not)
The apparent contradiction between "semaglutide reduces depression risk by 44%" and "semaglutide does not treat depression in an RCT" dissolves when you recognize they are measuring different things:
The observational studies measure whether semaglutide users, in the real world, experience less psychiatric deterioration over months to years. This captures the full package: weight loss, metabolic improvement, reduced food preoccupation, improved physical function, and any direct neurological effects.
The RCT measured whether semaglutide reduces depression symptom scores over 16 weeks in people who already have major depressive disorder. This tests whether semaglutide is an antidepressant. It is not, based on one trial.
The case reports capture individuals whose specific biology, psychology, or circumstances produce a negative response. No drug has a 100% positive response rate, and identifying who is at risk requires data we do not yet have.
The honest synthesis: semaglutide likely improves mental health outcomes for most users through indirect mechanisms (weight loss, metabolic improvement, reduced food obsession), does not appear to directly treat depression as a pharmacotherapy, and can worsen mood in a minority of individuals through mechanisms that are not yet well characterized.
For readers interested in how GLP-1 drugs affect alcohol use and other compulsive behaviors, the reward-system modulation may overlap with mood effects.
The Bottom Line
Large observational studies consistently show that semaglutide use is associated with lower rates of worsening depression, with the largest study to date (95,490 patients) reporting a 44% risk reduction. Randomized controlled trials confirm psychiatric safety but do not show direct antidepressant effects. Case reports document worsened depression in individual patients. The most likely explanation is that semaglutide improves mood indirectly through weight loss and metabolic improvement in most users, while a minority experience negative psychological effects through mechanisms that remain poorly understood.