The GLP-1 semaglutide is associated with a lower risk for psychiatric hospitalization and relapse in patients with bipolar disorder (BD) and diabetes and/or obesity.

In a large observational study, semaglutide was associated with a 21% lower risk for psychiatric hospitalization and a 17% lower risk for hospitalization for BD relapse.

Neither liraglutide nor dulaglutide was associated with a significant reduction in psychiatric hospitalization, although the investigators noted that fewer users of these drugs may have limited the analyses. When GLP-1s were evaluated as a group, however, their use was associated with a lower hospitalization risk in this patient population.

The findings raise the possibility that these medications could eventually be used as an adjunct to mood stabilizers in patients with BD and obesity or type 2 diabetes, said study investigator Mark Taylor, MD, professor at School of Medicine and Dentistry, Griffith University, in Gold Coast, Australia, and a consultant psychiatrist.

“Contrary to initial reports, these GLP-1s appear safe from a psychiatric perspective,” Taylor told Medscape Medical News.

The study was published online in Acta Psychiatrica Scandinavica.

Exploring the Metabolic-Psychiatric Connection

GLP-1 receptor agonists (RAs) are established treatments for diabetes and obesity, but interest in their potential psychiatric effects is growing.

Prior research has linked their use to psychiatric outcomes including lower rates of depression, anxiety, self-harm, and substance use, but their potential effects in BD remain largely unexplored.

In addition, obesity and type 2 diabetes are common comorbidities in BD, and recurrent psychiatric episodes frequently lead to hospitalization.

To examine the risk for psychiatric hospitalization and sick leave, investigators used Swedish national health registers to identify individuals with BD who were also receiving antidiabetic medications.

The study included 14,694 patients (mean age, 53.9 years; 60.5% women) who received antidiabetic medication between 2009 and 2024. During a mean follow-up of 6 years, 5200 patients used a GLP-1.

The study’s primary outcome was psychiatric hospitalization for any reason. The investigators also assessed hospitalizations from BD relapse and psychiatric sick leave lasting more than 14 days.

The investigators used a within-individual design that allowed each participant to serve as their own control. They compared periods of GLP-1 use with periods of nonuse, adjusting for time-varying factors including mood stabilizers, antipsychotics, antidepressants, medications for substance use disorders, benzodiazepines, and other antidiabetic medications.

Reduced Hospitalization Risk

During follow-up, 5288 participants experienced at least one psychiatric hospitalization.

Semaglutide use was associated with a 21% lower risk for psychiatric hospitalization than periods when the same patients were not using a GLP-1 (adjusted hazard ratio (aHR), 0.79; 95% CI, 0.69-0.91). When GLP-1s were analyzed as a group, the use was associated with a 13% lower risk (aHR, 0.87; 95% CI, 0.79-0.97).

Neither liraglutide nor dulaglutide was individually associated with a significant reduction in hospitalization risk, although these drugs were used less frequently, limiting statistical power.

Among 1274 participants who used both semaglutide and another GLP-1 during follow-up, semaglutide was associated with a 17% lower hospitalization risk than the other GLP-1s (aHR, 0.83; 95% CI, 0.71-0.98).

The finding was similar after excluding hospitalizations primarily related to substance use disorders, with a 20% lower risk (aHR, 0.80; 95% CI, 0.67-0.94).

Additional analyses yielded similar findings. After the first 30 days of use and nonuse were excluded, semaglutide remained associated with a 22% lower risk for psychiatric hospitalization (aHR, 0.78; 95% CI, 0.68-0.90).

The association also persisted after hospitalizations primarily related to substance use disorders were excluded, with a 20% lower risk (aHR, 0.80; 95% CI, 0.67-0.94).

Semaglutide showed a similar association with hospitalization for BD relapse. Among 1634 participants who experienced this outcome, semaglutide was associated with a 17% lower risk (aHR, 0.83; 95% CI, 0.69-0.99), whereas GLP-1 use as a group was associated with a 15% lower risk (aHR, 0.85; 95% CI, 0.74-0.97).

This was not true for psychiatric sick leave. Among the 10,526 participants included in this analysis, 2466 had medically certified psychiatric sick leave, but neither individual GLP-1s nor the drug class as a whole was associated with a significant reduction in this outcome.

The investigators proposed several potential explanations for the findings, including direct effects of GLP-1 signaling on brain function and indirect effects through weight loss.

“We don’t fully understand the direct brain effects of these agents, but my impression is that they may have a differing impact on neuroinflammation and the HPA [hypothalamic-pituitary-adrenal] axis,” Taylor said.

The reason for the apparent difference between semaglutide and other GLP-1s remains unclear. “Meta-analysis elsewhere indicates that semaglutide causes more weight loss than other GLP-1 RAs, so perhaps it’s more potent,” Taylor suggested.

The researchers noted that residual confounding remains a limitation, including the possibility that changes in psychiatric symptoms influenced whether patients initiated or continued semaglutide treatment.

“It is conceivable that worsening psychiatric symptoms may have reduced the likelihood of an individual being prescribed semaglutide or continued in people who were already well engaged with care providers,” they wrote.

The study did not include individual data on weight, A1c, or BMI, so the investigators were unable to determine whether changes in metabolic health contributed to the results.

The researchers emphasized the need for further study, ideally in randomized controlled trials of GLP-1s in people with BD.

Taylor said future randomized trials should assess outcomes beyond hospitalization alone.

“Improvements in symptom control, hospitalization rate, and death rate would be important outcomes to study,” he said.

A Signal Worth Pursuing

The potential neuropsychiatric effects of GLP-1s remain an area of interest for clinicians and investigators, Steve Strakowski, MD, professor, and vice chair of research in the Department of Psychiatry at Indiana University School of Medicine, Indianapolis, who was not involved in the research told Medscape Medical News.

“These findings are interesting as, given the hedonic effects of these compounds, clinicians and investigators have been wondering if they might impact the course of mood disorders, including BD,” Strakowski said.

However, he cautioned that large exploratory studies like this one can identify signals that are difficult to replicate or specifically link to treatment response.

“A single report does not mean semaglutide is ready for widespread treatment application for bipolar disorder,” he said.

The apparent difference between semaglutide and other GLP-1s was also less definitive. Although semaglutide was associated with lower hospitalization risk while liraglutide and dulaglutide were not, “the confidence intervals across the compounds significantly overlap,” Strakowski said.

He noted that unmeasured confounding could contribute to the observed differences. Nevertheless, he added, the findings support further investigation of GLP-1 RAs in BD.

“This early signal suggests some head-to-head studies could be informative and minor differences in mechanisms explored. Direct clinical trials are necessary to establish treatment response,” Strakowski added. 

Disclosure information for study authors is available in the original study publication. Strakowski reported having no relevant financial disclosures.