Daniel was 36 when the voices finally went quiet. It took multiple antipsychotic trials, but he’s finally thinking about going back to work. However, in the mirror he looks quite different, his trousers don’t fit, and he is now diabetic. He was warned that this might happen, and tried to adjust his diet and exercise, but it hardly worked.
Daniel represents many people with schizophrenia spectrum disorders. Ian, a lived experience group member, put it plainly:
It’s not me putting on weight, it’s the medication that’s making me put on weight (McPin Foundation, 2025).
Antipsychotics are the cornerstone of treatment for schizophrenia, and though clozapine and olanzapine are among the most effective, they are associated with weight gain and metabolic harm (Pillinger et al., 2020). And the cost goes well beyond appearance. People with severe mental illness have a 54% higher risk of coronary heart disease and an 85% higher risk of cardiovascular death than the general population (Correll et al., 2017). Weight gain also adds a second stigma to the condition (Gerlinger et al., 2013), pushing people to choose between their mental and physical health. Christopher, prescribed olanzapine after a manic episode, described the same bind:
I wanted to stop taking the medication but I knew that would make me worse too (YoungMinds, 2020)
For years, the answer was disappointing: some lifestyle advice, maybe metformin (Tonini E , 2026), and hope. Then glucagon-like peptide-1 (GLP-1) receptor agonists like semaglutide (brand: Ozempic; Snowden J, 2026) appeared. The question then became: do we know what drugs could help patients lose weight after receiving antipsychotic treatment? Stogios et al., (2026) tried to answer exactly that.
Methods
This systematic review and network meta-analysis of metabolic interventions in antipsychotic-treated patients with schizophrenia spectrum disorders (SSDs) was pre-registered on PROSPERO (CRD42024610768) and seven databases were searched, including unpublished registry data and non-English papers. Adult randomised clinical trials (RCT) were included, regardless of illness length, stage, state or treatment setting. Studies where >50% of patients had a non-schizophrenia-related diagnosis were excluded.
The authors used an “automation tool” (slightly vague) to exclude some records, and further screened the remaining records by hand. All screening, data extraction and bias assessment was completed by two independent people, with disagreements resolved by discussion among the wider author group.
Eligible studies were included in the frequentist random-effects network meta-analysis, with the primary focus on weight mean difference (kg). Weight loss treatments were compared against placebo or ‘usual care’. Certainty of evidence was rated using the Confidence in Network Meta-Analysis (CINeMA) tool.
Results
A whopping 18,498 records were found! Of these, 9,162 were removed before screening (5,036 flagged by the automation tool and 4,126 duplicates), leaving 9,336 to be screened by hand. This was whittled down to 95 eligible RCTs from 1965 to 2025 that examined 39 drugs in 5,898 patients. Ten trials were only narratively described: 3 had no good data, 7 were drugs no longer marketed.
Curiously, there were only 69 RCTs and 28 drugs that contributed to the primary outcome of average weight difference. Five drugs reduced weight compared with placebo, with moderate certainty of evidence: semaglutide (-10.98 kg, 95% confidence interval (CI) -13.33 to -8.62; 3 trials), liraglutide (-5.43 kg; 2 trials), topiramate (-3.95 kg; 5 trials), metformin (-3.86 kg; 16 trials), and exenatide (-2.97 kg; 3 trials). Semaglutide showed the largest weight reduction, while metformin was supported by the most trials. Three other drugs (ramelteon, nizatidine, and aripiprazole) also showed some benefit, but the certainty of evidence was very low.
The authors also attempted to quantify clinical significance by reporting “clinically meaningful weight loss”. This was defined as 5% or more weight loss from baseline and was compared to placebo. This was however only performed for 8 studies which had adequate data. Only semaglutide and metformin showed statistically significant benefits. Semaglutide had an odds ratio (OR) of 210.00 (95% CI 3.97 to 11,115.99; 1 trial), while metformin had an OR of 10.39 (95% CI 1.81 to 59.54; 4 trials). In other words, patients taking metformin had approximately 10 times the odds of achieving clinically meaningful weight loss compared with placebo or usual care.
The authors also ran sensitivity analyses (for example, removing studies at high risk of bias), and the five main drugs held up, although ramelteon and nizatidine dropped out. They also ran subgroup analyses. When they examined only “long-term” trials (at least 24 weeks of treatment), the effects of semaglutide (k=3) and metformin (k=6) held up. In clozapine-only treated patients, semaglutide (k=1) had the strongest mean weight difference, followed by exenatide (k=1). In olanzapine-only treated patients, there was no strong evidence to suggest that any drug had an effect as compared to placebo.
Conclusions
The authors concluded that among all the 39 drugs examined, semaglutide, liraglutide, exenatide, metformin and topiramate were the most effective, in people with schizophrenia spectrum disorders. Metformin was studied the most.
The authors also emphasized the need for adequately-powered and longer trials, more direct head-to-head comparisons between drugs (not just versus placebo), and inclusion of newer drugs like tirzepatide.
Strengths and limitations
The authors have invested an incredible amount of time and effort to answer a really important challenge that patients with SSDs have to deal with. This was done comprehensively (searching multiple databases, even including non-English studies to ensure wider capture), and robustly (duplicate screening, data extraction, risk of bias assessment, with multiple sensitivity and subgroup analyses). More recently published tools such as CINeMA and RoB2 were also implemented. They also included (relatively) newer drugs such as semaglutide, which could be compared indirectly to other drugs that had been on the market for a longer period of time. Effects were reported at the product-level (rather than at a drug class-level) which allows for clinicians and patients to understand which drug might work best.
While the findings of the study have shown interesting trends, I remain cautious about the interpretation of its results. This is for a number of reasons:
Firstly, many of the reported effects at the product-level are only supported by a small number of RCTs. Furthermore, the individual trials are characteristically very distinct: different countries, outpatient/inpatient setting, length of treatment, dose, illness stage or chronicity of illness, diagnostic criteria for SSDs etc. This means that the effects reported could be unique to a specific population with a certain ethnicity, severity of illness, at a specific dose and length of treatment of the drug. This makes it difficult for clinicians to know whether the findings apply to their patients.
Secondly, while it is understandable the authors had to feasibly exclude irrelevant studies, the validity of the ‘automation tool’ that they used during screening is unclear.
Thirdly, the inclusion of studies extending back to 1965 could increase the amount of available evidence, but treatment practices, patient populations and lifestyles may have changed considerably since then, potentially limiting the relevance of older findings to current practice.
Finally, it may be important to consider what is ‘clinically meaningful’ weight loss for the patient, which could go beyond the standard “5%” threshold used here. There could be two perspectives:
- What is the threshold needed to reduce cardiovascular risk?
- What is the weight loss desired by the patient? Is this the weight before antipsychotics were started?
Implications for practice
My personal take is that this study has further solidified that metformin can contribute to some weight loss in patients with SSDs who are currently taking antipsychotics. However, the extent of weight loss and whether that’s meaningful to the patient physiologically or psychologically (reducing the stigma associated with weight gain) remains to be seen. Semaglutide seems to show promise, and should be examined in greater detail in future work. Who would benefit the most? Is it safe to use for long periods of time, and how does it impact other psychiatric outcomes (e.g., relapse, symptom scores, suicidality)? Are there antipsychotic-specific combinations that work better (i.e., why is there no strong evidence for weight loss in olanzapine users)?
For instance, a known side effect of clozapine is slowing of gut movements (Every-Palmer S. & Ellis P. M., 2017). Can semaglutide, which is also known to have gastrointestinal side effects, worsen that? (Wharton et al., 2022). This would need a more thorough investigation before it can be routinely recommended in practice.
Also, as the authors have highlighted, as an important nuance in the article is what is being examined: the effects of starting weight loss drugs for weight gain that has already occurred following antipsychotic use. This is different from starting weight loss drugs to prevent weight gain when starting antipsychotics (Agarwal S. M. et al., 2022). The ideal situation for Daniel would be the latter, which is currently the focus of many ongoing and planned studies (more to come soon!).
Statement of interests
Aaron Yap (he/him) is funded by the Wellcome Trust and has no relevant conflicts of interest to declare. Claude was used to summarise the study in initial phases of planning the content of this article, but all opinions reflected here are his own.
Editor
Edited by Éimear Foley. ChatGPT assisted with language refinement and formatting during the editorial phase.
Links
Primary paper
Nicolette Stogios, Sri Mahavir Agarwal, Kateryna Maksyutynska, Halima Faisal, Lora Lee Pless, Toby Pillinger, Bailey Humber, Valerie H. Taylor, Gary Remington, Guy E. J. Faulkner, Margaret K. Hahn (2026) Pharmacological interventions for weight reduction in patients with schizophrenia treated with antipsychotics: a systematic review and network meta-analysis. JAMA Psychiatry 2026 83(9) 912-920.
Other references
Agarwal SM, Stogios N, Ahsan ZA. et al (2022) Pharmacological interventions for prevention of weight gain in people with schizophrenia. Cochrane Database Syst Rev 2022 10(10) CD013337.
Correll CU, Solmi M, Veronese N. et al (2017) Prevalence, incidence and mortality from cardiovascular disease in patients with pooled and specific severe mental illness: a large-scale meta-analysis of 3,211,768 patients and 113,383,368 controls. World Psychiatry 2017 16(2) 163-180.
Every-Palmer S, Ellis PM (2017) Clozapine-induced gastrointestinal hypomotility: a 22-year bi-national pharmacovigilance study of serious or fatal ‘slow gut’ reactions, and comparison with international drug safety advice. CNS Drugs 2017 31(8) 699-709.
Gerlinger G, Hauser M, De Hert M. et al (2013) Personal stigma in schizophrenia spectrum disorders: a systematic review of prevalence rates, correlates, impact and interventions. World Psychiatry 2013 12(2) 155-164.
McPin Foundation (2025) Antipsychotics & weight gain: ‘It’s not me, it’s the medication’. McPin website, 10 April 2025.
Pillinger T, McCutcheon RA, Vano L. et al (2020) Comparative effects of 18 antipsychotics on metabolic function in patients with schizophrenia, predictors of metabolic dysregulation, and association with psychopathology: a systematic review and network meta-analysis. Lancet Psychiatry 2020 7(1) 64-77.
Wharton S, Calanna S, Davies M. et al (2022) Gastrointestinal tolerability of once-weekly semaglutide 2.4 mg in adults with overweight or obesity, and the relationship between gastrointestinal adverse events and weight loss. Diabetes Obes Metab 2022 24(1) 94-105.
YoungMinds (2020) My experience of antipsychotics and weight gain. YoungMinds website, 31 January 2020.