Semaglutide and Bipolar Disorder: The Data

A nationwide Swedish cohort study in Acta Psychiatrica Scandinavica found that people with bipolar disorder had a lower rate of psychiatric hospitalization during the months they took semaglutide, compared with months they did not^[1]. The reduction was 21% for any psychiatric hospitalization and 17% for a bipolar-relapse admission specifically. The same dataset also tracked two related drugs, liraglutide and dulaglutide, and neither showed that individual pattern.

A letter published alongside the study makes a narrower and more careful point than the headlines: this split, one GLP-1 drug moving the outcome while its close relatives did not, needs a mechanism specific to semaglutide rather than an explanation that fits the whole drug class^[2]. That distinction matters more than the topline number. Here is what the underlying study measured, why the drug-by-drug difference is the real finding, and what is still unresolved.

What the new study found

Researchers including Heidi Taipale and Mark Taylor used Swedish national health registers covering 2009 through 2024 to identify 14,694 people with a bipolar disorder diagnosis who had also been prescribed a diabetes medication. The study did not compare those patients to a separate group. It compared each person's own hospitalization rate during periods on a GLP-1 receptor agonist (a drug class that mimics a gut hormone involved in blood sugar and appetite regulation) against that same person's rate during periods off one. Weighing a patient against their own history, not against a different set of patients, partly controls for the fact that people prescribed a newer diabetes drug are not a random slice of the bipolar-disorder population.

Of the 5,200 people who used a GLP-1 receptor agonist at some point, semaglutide use specifically was tied to a 21% lower rate of any psychiatric hospitalization (adjusted hazard ratio 0.79, 95% CI 0.69 to 0.91) and a 17% lower rate of hospitalization for a bipolar relapse (adjusted hazard ratio 0.83, 95% CI 0.69 to 0.99)^[1]. Across all three GLP-1 drugs combined, the reduction was smaller, 13% (adjusted hazard ratio 0.87, 95% CI 0.79 to 0.97), because liraglutide and dulaglutide diluted the semaglutide-only result rather than reinforcing it^[1].

Why the drug-by-drug split is the real story

If the lower hospitalization rate came from something every GLP-1 drug shares, better blood sugar control, weight loss, fewer physical health crises that can destabilize mood, you would expect liraglutide and dulaglutide users to show a similar pattern. They did not^[1]. The letter, titled "Semaglutide in Bipolar Disorder: A Differential Finding Needs a Differential Mechanism," argues this is the actual result worth explaining^[2]. The study authors raise possible semaglutide-specific mechanisms, including a drug effect on neuroinflammation (immune-driven inflammation inside the brain, distinct from inflammation elsewhere in the body) and neurotrophic effects (support for the growth and survival of brain cells). They are careful to frame these as candidate explanations, not established ones^[1].

What we don't yet know

Two things temper this finding. First, the study is observational. It cannot rule out that people who stayed on semaglutide longer were already on a better trajectory for reasons the data can't see, a limit the authors state directly: causality cannot be attributed in an observational study^[1]. Second, the psychiatric-safety picture for semaglutide is not uniformly reassuring in one direction. A 2025 case report in The Journal of Clinical Psychiatry described a 77-year-old woman with bipolar II disorder who experienced brief, pleasant euphoria after her first two semaglutide doses, an effect the authors called the first documented case of GLP-1-induced acute euphoria^[4].

One case report and one large cohort finding are not a contradiction. A drug can lower average hospitalization risk across thousands of patients while still producing an unusual mood reaction in an individual. But the 2025 case report is evidence that semaglutide's effects on mood in bipolar disorder are not fully mapped in either direction, and a single nationwide dataset, however well designed, is not the last word.

Regulators are already watching GLP-1 drugs and mental health

This is not the first time GLP-1 receptor agonists have drawn psychiatric scrutiny. In 2023 the European Medicines Agency opened a review of semaglutide, liraglutide and related drugs after case reports of suicidal thoughts, and its Pharmacovigilance Risk Assessment Committee concluded in April 2024 that the available evidence does not support a causal association between the drug class and suicidal or self-injurious thoughts^[3]. That conclusion answers a different question, suicidality, not bipolar-relapse risk, but it shows regulators are actively monitoring the psychiatric effects of this drug class instead of treating it as psychiatrically inert.

What this means if you have bipolar disorder

None of this changes the label. Semaglutide, sold as Ozempic, Wegovy and Rybelsus, is approved for type-2 diabetes and weight management, not for bipolar disorder, and nothing here is a reason to start or stop it for a mood-related purpose. If you have bipolar disorder and are already taking, or considering, semaglutide for diabetes or weight management, the useful step is telling your psychiatrist and your prescribing clinician about both this finding and the case-report signal, so mood changes in either direction get tracked rather than dismissed. A randomized trial, not another cohort study, is what would actually settle whether semaglutide affects bipolar disorder outcomes and through what mechanism.