The Semaglutide-Only Signal the Bipolar Headlines Will Flatten

A drug class becomes famous, and every new headline about it gets written as if the whole class moved together. That is how a single-molecule signal in bipolar disorder becomes "GLP-1s may help" before anyone checks whether the other molecules in the class did the same thing.

The Griffith University study of semaglutide (sold as Ozempic and Wegovy), a within-person comparison across 15 years of Swedish nationwide data covering nearly 15,000 adults with bipolar disorder, isolates exactly that question. It found semaglutide was associated with a 21% lower risk of psychiatric hospitalization during periods when patients were taking it. Other GLP-1 drugs in the same dataset did not produce that signal. The null is the finding.

The proposed mechanism is hypothesis, not measurement. Researchers suggest GLP-1 signaling may dampen neuroinflammation or cellular stress, and that semaglutide may do something the other molecules in the class do not. Until a randomized trial compares semaglutide head-to-head with another GLP-1 in this population, the question stays open.

The press cycle will frame this as "GLP-1s for bipolar." The data says "semaglutide, specifically, in a within-person comparison that controls for some confounders but not all." That gap is what a randomized trial is for. Until one lands, no clinician should change a prescription and no patient should ask for one.