People living with schizophrenia spectrum disorders carry a disproportionately heavy metabolic burden. Rates of obesity, elevated blood sugar, and related complications run significantly higher in this group than in the general population, and that gap translates into real differences in life expectancy. Several factors drive this: the disorders themselves alter appetite and activity patterns, and many medications prescribed to manage psychiatric symptoms are known to promote weight gain and raise blood glucose over time.
Against that backdrop, a class of compounds called glucagon-like peptide-1 receptor agonists, or GLP-1 RAs, has attracted considerable scientific attention. These peptides mimic a gut hormone that influences insulin release, appetite signalling, and the rate at which the stomach empties. In broad-population trials they have produced meaningful reductions in body weight and blood sugar. The open question was whether those effects would hold, and whether the safety profile would differ, in people with schizophrenia spectrum disorders specifically.
A systematic review and meta-analysis published in BJPsych Open in 2026 set out to answer that question by pooling every placebo-controlled randomised controlled trial that had tested the GLP-1 peptide semaglutide in this population. The analysis, led by Mike Trott and colleagues, identified three qualifying trials and synthesised their results. What follows is a plain-English walkthrough of what the researchers found.
Why this population needs dedicated research
General-population findings cannot always be transplanted to people with schizophrenia spectrum disorders. These individuals often take antipsychotic medications that directly affect metabolic pathways, alter hunger hormones, and promote fat storage. Their baseline physiology, medication interactions, and adherence patterns can all differ from trial populations used to establish standard dosing guidance.
Because of this, researchers and clinicians have called for sub-group-specific evidence rather than extrapolating from broader datasets. The 2026 meta-analysis represents one of the first attempts to gather and formally pool what randomised evidence does exist, rather than relying on case reports or observational data.
What the trials actually tested
The research team conducted a preregistered search and identified three randomised controlled trials that met their criteria. All three tested semaglutide against placebo in adults with schizophrenia spectrum disorders, using doses ranging from 1.0 to 2.0 milligrams. Trial durations ran between 26 and 36 weeks, giving a follow-up window of roughly six to nine months. Across the three studies, 258 participants were included in the pooled analyses.
Notably, the researchers had also planned to capture trials of another dual-acting GLP-1 and GIP receptor agonist peptide. No qualifying trials for that compound in this population existed at the time of the review, so all quantitative results come from the semaglutide data alone.
Outcomes were pooled using a random-effects meta-analysis model, which is designed to account for genuine variation in results across different study settings and populations. The certainty of the evidence was graded using GRADE criteria, a widely used framework for rating how confident researchers can be in a pooled finding.
Body weight and blood sugar findings
On the primary metabolic outcomes, the pooled results were substantial. Participants assigned to semaglutide lost an average of 11.32 kilograms more than those on placebo, with a 95 percent confidence interval ranging from 7.29 to 15.35 kilograms. Body mass index fell by an average of 3.58 units more in the semaglutide group.
Blood sugar markers also shifted. Haemoglobin A1c, a measure of average blood glucose over roughly three months, fell by 0.37 percentage points compared to placebo. Fasting glucose dropped by 0.54 millimoles per litre. Both of these changes were statistically significant, meaning they are unlikely to be explained by chance variation across the small number of trials.
The researchers described these changes as clinically meaningful, a phrase that signals the magnitude was large enough to plausibly matter in a real-world sense, not merely reach statistical thresholds. In a population where metabolic risk is elevated at baseline, reductions of this scale represent a notable finding, though the authors were careful to note the evidence base is still small.
Adverse events observed
Any honest assessment of a compound's research profile has to include what went wrong as well as what improved. The adverse event analysis found that semaglutide was associated with meaningfully higher rates of three gastrointestinal symptoms compared to placebo.
Abdominal pain occurred at a rate roughly 2.93 times higher in the semaglutide group. Vomiting was approximately 2.57 times more common. Constipation was about 3.23 times more frequent. These elevated risks are consistent with the known pharmacology of GLP-1 receptor agonists in other populations. The stomach-slowing effect that contributes to reduced appetite also tends to produce these gastrointestinal complaints, particularly during dose escalation.
Crucially, the analysis found no evidence of increased risk of serious adverse events in people with schizophrenia spectrum disorders taking semaglutide. The psychiatric safety signal that some might expect to see, given that the participants were managing an existing psychiatric condition, did not emerge in the pooled data. The authors acknowledge the trials were not powered to detect rare serious events with confidence, and this is an area that longer, larger studies will need to address.
Limitations and the call for more research
Three trials totalling 258 participants is a thin evidence base from which to draw firm conclusions. The authors were transparent about this. The trials varied in length, dosing, and population characteristics, and the random-effects model can only partially account for that heterogeneity.
The longest trial ran 36 weeks. That window is enough to observe short-term metabolic changes but is insufficient to evaluate what happens when treatment continues for years, which is the clinical reality for most people managing a schizophrenia spectrum disorder. Long-term cardiovascular outcomes, weight regain after discontinuation, and effects on psychiatric symptom scores were not robustly assessable from the existing data.
The review team explicitly called for larger and longer trials, with particular emphasis on the absence of any randomised evidence for the dual GLP-1 and GIP receptor agonist peptide in this population. They framed their findings as supportive of semaglutide as a promising adjunctive metabolic intervention, while making clear that the evidence remains preliminary and that characterising the full safety and efficacy picture will require considerably more research.
Broader context for GLP-1 peptide research
The 2026 meta-analysis sits within a much larger and rapidly expanding body of literature on GLP-1 receptor agonist peptides. Research in the general population has documented not only weight and glycaemic effects but also signals around cardiovascular risk, kidney function, and neurological outcomes. Whether any of those downstream effects extend to people with schizophrenia spectrum disorders in similar or modified form is an open scientific question.
What this particular analysis contributes is a formal synthesis of the available randomised evidence for a population that has historically been underrepresented in metabolic research. The findings do not resolve the clinical questions, but they mark a concrete step toward a literature that addresses the specific needs of people whose metabolic risk is compounded by both their diagnosis and its treatment.
