GLP-1 drugs in schizophrenia: 7 trials

Seven trials met the review's criteria. Together, they included 446 adults with schizophrenia spectrum disorders who were receiving antipsychotic treatment. The systematic review, published online on 27 July 2026, found that GLP-1 receptor agonists were associated with lower body weight, body mass index, waist circumference and fasting plasma glucose than control^[1]. The spread between trials was extreme, so the pooled weight estimate needs careful reading.

The pooled result

The review covered semaglutide and other medicines that activate the glucagon-like peptide-1 receptor, a signaling pathway involved in blood glucose control and appetite. Its authors searched electronic databases from inception to their final search date, restricted the analysis to randomized controlled trials and assessed bias with the Cochrane Risk of Bias 2 tool. They also used GRADE, a framework for judging how certain a body of evidence is, and had registered the protocol in PROSPERO before completing the review^[1].

For body weight, the pooled difference was -6.10, with a 95% confidence interval from -12.10 to -0.10 and a P value of 0.05. The PubMed abstract does not state the unit attached to that estimate, so we are not adding one. A confidence interval gives the range of effects compatible with the analysis under its assumptions. Here, the upper boundary sat close to no difference. The review also reported statistically significant reductions in body mass index, waist circumference and fasting plasma glucose^[1].

Results for blood lipids, insulin-related measures and blood pressure were inconsistent and generally did not reach statistical significance. Serious adverse events were less frequent in the intervention groups, but that finding comes from a pool of only 446 participants and does not settle uncommon safety risks. The authors rated the certainty of evidence from low to moderate, mainly because the estimates were imprecise and the trial results varied^[1].

Why the 99% heterogeneity matters

The body-weight analysis had an I-squared value of 99%^[1]. Results diverged sharply. I-squared estimates how much of the variation between study results is beyond what chance alone would be expected to produce, and a value this high means the trials were not giving one tightly consistent answer. Different GLP-1 medicines, participant groups, antipsychotic regimens, treatment periods and study designs may all contribute. The abstract does not quantify which factor explains it.

The pooled direction favors treatment, but the estimate is imprecise. One average cannot be carried cleanly into a clinic or applied to an individual patient when the component results differ this much. A class-wide estimate can also hide differences between older GLP-1 drugs and semaglutide, as well as differences between people taking clozapine, olanzapine or other antipsychotics. The larger recent semaglutide trials put more specific numbers on the question.

What the semaglutide trials add

The HISTORI trial randomized 154 adults with schizophrenia, prediabetes and overweight or obesity to semaglutide or placebo for 30 weeks. Semaglutide lowered HbA1c, a measure of average blood glucose, by 0.46 percentage points and body weight by 9.21 kg relative to placebo. It did not significantly change the brief schizophrenia symptom score used by the investigators. Gastrointestinal symptoms were more frequent, while serious adverse-event counts did not differ between groups^[2].

COaST asked a narrower question in adults with schizophrenia or schizoaffective disorder who were taking clozapine and had obesity. Only 31 participants were randomized because recruitment stopped early when the study drug became unavailable. At week 36, semaglutide produced a 13.46 percentage-point greater body-weight reduction than placebo. The trial found no difference in psychosis symptoms or clozapine and norclozapine blood concentrations, and no serious adverse events were judged related to treatment^[3]. Its small sample makes it encouraging evidence, not a final safety answer.

A second Danish trial enrolled 73 people with schizophrenia spectrum disorders who had early blood-glucose abnormalities while taking clozapine or olanzapine. After 26 weeks, semaglutide reduced HbA1c by 0.25 percentage points, body weight by 9.2 kg, waist circumference by 7.0 cm and fat mass by 6.1 kg relative to placebo. Psychiatric adverse events were similar between groups, and gastrointestinal events were common but described as mild and temporary^[4].

What this evidence does not establish

The review cannot show that every GLP-1 receptor agonist works equally well in every person taking an antipsychotic, establish long-term cardiovascular benefit in schizophrenia or rule out uncommon adverse events. Its pooled safety comparison is especially limited by the modest total sample. The recent semaglutide trials ran for 26 to 36 weeks. They measured changes in weight and glucose, but they cannot answer years-long questions about durability, cardiovascular events or psychiatric course.

The results also do not support changing psychiatric medication without clinical supervision. Antipsychotics treat serious illnesses, and stopping or altering them can carry immediate risk. Adding a prescription metabolic medicine creates a second treatment that needs its own monitoring. The trials studied GLP-1 drugs as adjuncts to continuing antipsychotic care, not as substitutes for it.

Why this matters

Antipsychotic-associated weight gain and glucose problems are often treated as unavoidable trade-offs. These randomized data give clinicians a stronger basis for discussing a GLP-1 receptor agonist with selected adults while keeping the limits visible. The pooled result points in a favorable direction, and the recent semaglutide trials provide more specific estimates. Neither removes the need to follow psychiatric symptoms, glucose measures, weight and adverse effects together.

Semaglutide is a prescription medicine, and its approved uses and access rules vary by jurisdiction. Our semaglutide evidence page summarizes the wider evidence base, while the regulation section tracks country-level status. Anyone considering a GLP-1 medicine alongside an antipsychotic should involve the clinician managing the psychiatric treatment rather than changing either medicine independently.