A preregistered systematic review of glucagon-like peptide-1 receptor agonists in schizophrenia spectrum disorders ↗, published May 25 in BJPsych Open, pooled three placebo-controlled randomised trials covering 258 adults and returned a mean body-weight loss of 11.32 kg (about 25 pounds, with a 95 percent confidence interval running from 7.29 to 15.35 kg) on semaglutide ↗ at 1.0 to 2.0 mg weekly over 26 to 36 weeks. The same preregistered search returned zero randomised trials of tirzepatide ↗, the dual GIP/GLP-1 agonist that has consistently outperformed semaglutide on weight in general-population obesity trials. The meta-analysts called for tirzepatide work specifically.
Schizophrenia spectrum disorders shorten life expectancy by 15 to 25 years, and cardiometabolic disease, driven heavily by antipsychotic-induced weight gain and the lifestyle correlates that come with severe mental illness, is the dominant contributor. The large semaglutide weight trials (STEP 1, STEP 8) and the tirzepatide weight trial SURMOUNT-1 all excluded patients with active psychotic illness as a matter of routine eligibility design. That exclusion is the reason this meta-analysis is the first of its kind, even though the population in question has more to gain from sustained weight reduction than almost any other clinical group.
The three included trials were small and short. Cohort sizes ranged from roughly 60 to 110 patients per trial, follow-up topped out at 36 weeks, and the semaglutide doses (1.0 to 2.0 mg weekly) sit in the Ozempic diabetes-management range rather than the 2.4 mg Wegovy weight-management ceiling. Despite those constraints, pooled outcomes moved on every metric the meta-analysts pulled. Body mass index fell by 3.58 kg/m² (95 percent CI minus 4.86 to minus 2.30), which is roughly the BMI gap between an obese category and an overweight category. Glycated hemoglobin (HbA1c, the three-month blood-sugar average) dropped 0.37 percentage points (95 percent CI minus 0.51 to minus 0.22). Fasting glucose dropped 0.54 mmol/L, about 10 mg/dL.
The adverse-event signature looked familiar to anyone who has read a STEP trial. Abdominal pain ran 2.93 times more common on semaglutide than on placebo (95 percent CI 1.13 to 7.60). Vomiting was 2.57 times more common (95 percent CI 1.39 to 4.77). Constipation was 3.23 times more common (95 percent CI 1.14 to 9.18). Serious adverse events did not separate from placebo. The authors flagged the gastrointestinal pattern as consistent with the known general-population profile of GLP-1 receptor agonists. The wide confidence intervals on every ratio reflect the small total sample.
The tirzepatide gap is the structural news. The drug is approved for obesity, has delivered roughly 50 percent more weight loss than semaglutide in general-population obesity (SURMOUNT-1 vs STEP 1, cross-trial), and is the natural next molecule to test in a group whose metabolic burden is so heavy. The preregistered search found nothing to pool. The reasons are partly mechanical (tirzepatide is newer, FDA-approved for obesity in 2023) and partly the same eligibility-design reflex that kept the semaglutide work small. Lilly's investigator-sponsored portfolio for SSDs populations is not visible in the published literature yet.
Translating to practice today, the meta-analysis supports semaglutide as an adjunctive metabolic intervention in schizophrenia spectrum disorders rather than as a standalone replacement for any psychiatric medication. The pooled cohort of 258 patients is small. The certainty of evidence under GRADE was assessed but the abstract does not flag any endpoint as high-certainty. What the meta does establish is that the gastrointestinal-tolerability story carries over from the general population into a group whose medication burden, cognitive symptoms, and primary-care access all complicate adherence in ways the registration trials never tested.
A larger and longer RCT, ideally with tirzepatide or with semaglutide at the 2.4 mg Wegovy ceiling dose, would close the comparison to the general-population effect magnitude (Wegovy's STEP 1 returned about 14.9 percent body-weight loss at 68 weeks, the canonical anchor for semaglutide weight efficacy). Until then, prescribers treating antipsychotic-associated weight gain have one preregistered meta to point to, with confidence intervals wide enough that the next trial could meaningfully revise the headline number in either direction.