Ozempic users with bipolar disorder had a 21% lower rate of psychiatric hospitalization in a 15,000-person study

The protective effect showed up with one drug and not the other two. People with bipolar disorder who took semaglutide, the compound sold as Ozempic and Wegovy, had a 21% lower rate of psychiatric hospitalization compared with periods when they were not on a GLP-1 medicine, according to a Swedish registry study covering nearly 15,000 people with the diagnosis. Liraglutide and dulaglutide, two other drugs in the same class, showed no comparable benefit in the same dataset. Griffith University’s Mark Taylor was among the researchers on the project, which drew on Sweden’s national health records rather than a newly recruited trial population.

A Swedish Registry Spanning Fifteen Years

The study, published in Acta Psychiatrica Scandinavica, tracked hospitalization rates for the same individuals across 2009 through 2024, comparing each person’s own record during stretches when they were filling a GLP-1 prescription against stretches when they were not. That within-person design controls for a lot of what usually complicates this kind of research, since it compares a patient to an earlier or later version of themselves rather than to a different person entirely, though it still cannot rule out that whatever prompted a doctor to prescribe semaglutide in the first place, rather than the drug itself, was doing some of the work.

One Drug Stood Out, Two Did Not

Semaglutide was the only medication in the trio to show a statistically meaningful association with fewer hospital stays. “People with bipolar disorder who were taking semaglutide had significantly lower rates of psychiatric hospitalisation compared with periods when they were not taking GLP-1 medicines,” Taylor said, describing results also summarized in the original study coverage. Liraglutide and dulaglutide, despite working through the same receptor pathway, did not produce a statistically significant drop in hospitalization during the periods patients used them, a distinction that argues against treating GLP-1 drugs as a single interchangeable class for psychiatric purposes.

Taylor framed the finding as suggestive rather than settled. “The findings add to growing evidence that GLP-1 receptor agonists may have benefits beyond diabetes and obesity treatment, and could represent a promising new avenue for bipolar research,” he said, a statement that stops short of recommending semaglutide as a psychiatric treatment and instead points toward the randomized trials that would be needed to confirm the pattern.

The distinction between the three drugs is itself a finding worth sitting with. Liraglutide reached the market well before semaglutide and has a longer track record in both diabetes and obesity care, while dulaglutide is prescribed almost exclusively for diabetes rather than weight management. If a GLP-1 drug’s psychiatric benefit tracked simply with how strongly it activates the shared receptor, all three would be expected to move together. Instead, only semaglutide separated from the pack in this cohort, which is the kind of result that tends to sharpen a research question rather than close it.

A Mechanism Still Confined to Mice

Why semaglutide specifically might lower hospitalization risk is not answered by the Swedish data, which tracked outcomes without explaining the biology behind them. A separate line of laboratory research offers one plausible thread: a mouse study published in Nature Communications found that semaglutide reduced inflammation in the brain by blocking immune cells from crossing into brain tissue and calming inflammatory activity in microglia, the brain’s resident immune cells. That mechanism, observed in animals rather than in people with bipolar disorder, is consistent with the idea that GLP-1 drugs could influence mood through inflammation rather than through any direct action on classic psychiatric pathways, but it has not been tested in the population the Swedish study examined.

Bipolar Disorder’s Global and Domestic Scale

The condition the study addresses is common enough that even a modest, unconfirmed effect would matter broadly. The World Health Organization estimates that about 36 million people worldwide live with bipolar disorder, roughly one in every 226 people. In the United States specifically, the National Institute of Mental Health puts past-year prevalence among adults at 2.8%, with a lifetime rate near 4.4%, and notes the condition is most common among adults aged 18 to 29 before declining with age.

That scale is part of why a Swedish national registry was even possible in the first place. A condition affecting close to three in a hundred adults in a given year generates enough hospitalization records, prescription fills and follow-up years to support a within-person comparison spanning a decade and a half, something a smaller or newly recruited study population could not have supplied on its own.

Such context leaves unchanged what the Swedish registry actually showed, which is an association observed in medical records, not a causal claim tested in a clinical trial designed for psychiatric outcomes. Taylor and his coauthors have not said whether a prospective trial testing semaglutide specifically for bipolar disorder is planned, leaving the 21% figure as a lead to follow rather than a treatment recommendation to act on.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview