Weight-Loss Drug Tames Blood Sugar in Schizophrenia Patients

Here’s something you don’t see often in psychiatry research: a treatment that actually works before the damage is done. Danish researchers gave semaglutide to 36 people with schizophrenia who were gaining weight and showing early blood sugar problems from their antipsychotic meds. Six months later, nearly half had pulled their metabolic markers back into the safe zone. On placebo? Three percent.

The timing matters. These weren’t patients who’d already developed diabetes after years on clozapine or olanzapine, the two antipsychotics most notorious for packing on pounds and wrecking metabolism. The researchers caught them early, within five years of starting treatment, when their hemoglobin A1c levels had crept into the danger zone but hadn’t crashed through to full-blown diabetes yet. That window between “uh oh” and “too late” is where this study lives.

Anyone who works in serious mental illness knows the brutal tradeoff: clozapine and olanzapine are sometimes the only drugs that work for psychosis, but they turn patients into metabolic wrecking balls. Weight shoots up 20, 30, 40 pounds. Blood sugar goes haywire. And people with schizophrenia already die 15 years younger than everyone else, mostly from heart disease. Switching to a “cleaner” antipsychotic often means psychotic symptoms come roaring back.

So you’re stuck managing the side effects. Metformin helps a little, maybe shaves off 2 to 6 kilograms if you’re lucky. Topiramate, same story. Nothing really moves the needle until now, apparently.

The Numbers Tell an Unusual Story

Semaglutide dropped hemoglobin A1c by 0.31% compared to a barely-there 0.02% increase on placebo. That’s a 0.25 percentage point difference, which sounds modest until you realize what it meant functionally: 12 of 28 people on the drug hit blood sugar levels below 5.4%, the threshold where diabetes risk really drops. One person out of 37 managed that on placebo.

Weight loss averaged 8.7 kilograms (about 19 pounds) in the semaglutide group versus essentially nothing on placebo. Waist circumference shrank 7 centimeters. The drug also carved off 5 kilograms of pure fat mass, though oddly the reduction in visceral fat, the dangerous belly fat wrapped around organs, didn’t quite reach statistical significance.

“Semaglutide was found to significantly reduce hemoglobin A1c level and body weight over a 26-week period,” the researchers reported in JAMA Psychiatry.

What’s interesting is how much individual variation showed up in the data. Some people responded dramatically, others barely budged. The study doesn’t dig into why, whether it’s genetic differences, medication interactions, baseline insulin resistance, or just random biological luck. That kind of heterogeneity usually means the drug works through multiple pathways, not just one simple mechanism.

The psychiatric symptoms stayed stable, which is frankly more important than the metabolic wins. Nobody wants a diabetes drug that destabilizes schizophrenia. Hospitalizations for psychiatric crises happened at similar rates in both groups. Symptom severity scores moved in parallel. The antipsychotics kept doing their job.

Side Effects Were Predictable, Except One Thing

Nausea hit 47% of the semaglutide group versus 41% on placebo, which seems high for placebo until you remember these are people taking clozapine or olanzapine, drugs that mess with your gut on their own. Vomiting, constipation, the usual GLP-1 agonist complaints. Four people quit because of gastrointestinal misery, all in the active treatment arm.

One patient died of sudden cardiac death a month into the trial. He was only on the 0.25 mg starter dose. Autopsy didn’t point to semaglutide, but sudden cardiac death is already elevated in schizophrenia patients, especially those on antipsychotics that prolong the QT interval. Did the drug contribute? Impossible to say from one case, but it’s the kind of signal that demands attention in larger trials.

The unexpected finding was nicotine. Among the roughly 40% of participants who smoked, which is actually lower than typical for schizophrenia populations, those on semaglutide showed reduced nicotine dependence scores. The effect was modest but consistent with emerging research suggesting GLP-1 drugs dampen reward pathways in the brain. They didn’t see similar effects for alcohol or other drugs, though the trial specifically excluded people with active substance use disorders, so there wasn’t much room to detect a signal there.

“Approximately one-half of individuals treated with semaglutide achieved low-risk hemoglobin A1c levels, compared with the placebo group,” the authors noted.

The study has the usual problems. Six months isn’t long enough to know if the benefits last or if people regain weight once they stop. The 1 mg weekly dose is half what’s approved for obesity treatment, so maybe a higher dose would work better, or maybe it would just cause more people to quit from side effects. The participants were young (average age 35), mostly White, relatively healthy aside from their psychiatric diagnosis and weight issues. Generalizing to older, sicker, more diverse populations is guesswork.

And then there’s cost. Semaglutide runs thousands of dollars a year. People with serious mental illness are disproportionately poor, uninsured or underinsured, and already navigating a fragmented healthcare system. The researchers suggest public health programs should subsidize GLP-1 drugs for high-risk psychiatric patients, which sounds reasonable until you imagine trying to get that through a budget committee.

Still, this is one of the first studies to show you can actually reverse early metabolic damage from antipsychotics rather than just slow it down. Whether that translates to fewer heart attacks and longer lives, nobody knows yet. But catching the problem before it becomes permanent feels like progress, assuming anyone can actually access the treatment.

JAMA Psychiatry: 10.1001/jamapsychiatry.2025.3639