Semaglutide users showed 42% fewer psychiatric hospitalizations in a study of nearly 100,000 people

University of Eastern Finland / Karolinska Institutet / Griffith University

The drugs were designed to control blood sugar and reduce weight. But something else kept showing up in the data. People taking GLP-1 receptor agonists - the class that includes semaglutide, sold as Ozempic and Wegovy - were spending less time in psychiatric hospitals. They were taking fewer sick days for mental health reasons. Their rates of substance use disorders were dropping.

A new study, one of the largest to examine this pattern, now quantifies the association with striking precision.

Tracking 100,000 people through 13 years of Swedish records

The research, published in The Lancet Psychiatry, drew on Swedish national registers covering the period from 2009 to 2022. It included nearly 100,000 participants, over 20,000 of whom had used GLP-1 medications. The study was a collaboration between the University of Eastern Finland, Karolinska Institutet in Stockholm, and Griffith University in Australia.

Rather than comparing GLP-1 users to non-users - an approach vulnerable to selection bias - the researchers used a within-individual design. They compared each person's outcomes during periods when they were taking GLP-1 medications against periods when they were not. This controls for stable individual characteristics like genetics, personality, and baseline mental health that might otherwise confound the results.

During periods of semaglutide use specifically, participants showed a 42% reduction in hospital care and sickness absence due to psychiatric reasons compared with their own non-use periods. For depression, the reduction was 44%. For anxiety disorders, 38%.

The substance use signal that surprised even the researchers

The psychiatric findings were notable, but the substance use data caught the research team off guard. Hospital care and sickness absence related to substance use disorders were 47% lower during periods of semaglutide use. GLP-1 receptor agonist use was also associated with a reduced risk of suicidal behavior.

Professor Mark Taylor from Griffith University, one of the study's authors, notes that the substance use finding was not entirely unexpected in direction - a previous Swedish registry study had already linked GLP-1 medications to reduced risk of alcohol use disorder, and alcohol problems often have downstream effects on mood and anxiety. But the magnitude of the association was larger than anticipated.

Research Director Markku Lahteenvuo from the University of Eastern Finland points to several possible mechanisms. Weight loss can improve body image and self-esteem. Better glycemic control in diabetes patients may relieve a chronic source of stress and anxiety. Reduced alcohol consumption removes a potent depressant and anxiety trigger. But there may also be direct neurobiological effects - GLP-1 receptors are found throughout the brain, including in reward circuitry, and the medications may alter how the brain processes reward and craving.

The honest answer is that we do not yet know which mechanism dominates, or whether several work in concert. This is a registry-based study. It can tell us that something is happening. It cannot tell us exactly why.

What registry data can and cannot establish

The strengths of this study are real. Nearly 100,000 participants. Thirteen years of follow-up. A within-individual design that controls for many sources of confounding. National registers that capture hospitalizations and sick leave comprehensively, without relying on self-report.

But registry studies have inherent limitations. The within-individual design controls for stable characteristics, but it cannot control for time-varying factors. If people start GLP-1 medications during a period when they are also making other health changes - exercising more, drinking less, engaging with healthcare more broadly - the medication gets credit for improvements that may partly reflect those concurrent changes.

The study also measures relatively severe outcomes: hospitalizations and sickness absence. These capture the serious end of psychiatric illness but miss milder symptoms. Someone whose anxiety improves from debilitating to merely unpleasant would not show up in hospital records. Conversely, someone whose depression worsens but never reaches the threshold for hospitalization would also be invisible to this methodology.

Previous research on GLP-1 medications and mental health has been inconsistent, though much of it relied on smaller samples. Some clinical trials have raised concerns about potential psychiatric side effects, including suicidal ideation, particularly with certain GLP-1 formulations. Regulatory agencies in multiple countries are monitoring these signals. This study's findings are encouraging, but they do not settle the question of psychiatric safety for all patients.

Finally, the population studied was Swedish, drawing on a healthcare system and cultural context that may not generalize globally. Prescribing patterns, patient populations, and access to mental health care differ substantially across countries.

The gap between association and prescription

The distance between these findings and a clinical recommendation is considerable. Showing that people who take semaglutide for diabetes or obesity also experience fewer psychiatric hospitalizations is not the same as showing that semaglutide should be prescribed for depression or anxiety. That would require randomized controlled trials designed specifically to test psychiatric outcomes as primary endpoints - trials that have not yet been conducted.

The metabolic-psychiatric connection is itself well established. Diabetes and obesity are associated with elevated risk of depression, anxiety, and substance use disorders. People with serious mental illness die 15 to 20 years earlier than the general population, largely from metabolic and cardiovascular disease. Any medication that effectively treats metabolic conditions may produce downstream psychiatric benefits simply by improving overall health.

What makes the GLP-1 findings intriguing is the possibility that the psychiatric benefits go beyond metabolic improvement - that the drugs may act directly on brain circuits involved in mood, anxiety, and reward. If that is the case, it would open a genuinely new avenue for psychiatric treatment. But proving it requires a different kind of evidence than this study provides.

For now, the data suggest that patients already taking GLP-1 medications for metabolic reasons may be getting an additional benefit they did not expect. Whether that benefit can be harnessed deliberately, for patients who need psychiatric help rather than metabolic help, remains one of the more interesting open questions in psychopharmacology.