Skip to main content
NCBI home page
Diabetes Care logoLink to Diabetes Care
. 2026 Mar 4;49(5):800–807. doi: 10.2337/dc25-2041

Semaglutide Effects on Insulin Sensitivity and β-Cell Function in Patients With Schizophrenia, Prediabetes, and Obesity Treated With Second-Generation Antipsychotics: Findings From the HISTORI Trial, a 30-Week Randomized, Placebo-Controlled Trial With Semaglutide 1.0 mg Weekly

Ashok A Ganeshalingam 1,2,3,✉, Nicolai Uhrenholt 2,4,5,6, Sidse Arnfred 6,7, Peter Gæde 8,9, Andreas K Pedersen 10, Niels Bilenberg 2,5, Jan Frystyk 1,2,3
PMCID: PMC13094869  PMID: 41778920

Abstract

OBJECTIVE

To examine the effects of semaglutide on insulin sensitivity, insulin resistance, and β-cell function and explore whether these changes were mediated by weight loss in overweight or obese individuals with schizophrenia and prediabetes receiving second-generation antipsychotics.

RESEARCH DESIGN AND METHODS

In this 30-week, double-blind trial, 154 participants were randomized to semaglutide (n = 77) or placebo (n = 77); 141 (91.5%) completed the study. Baseline and end-of-study assessments included fasting glucose, insulin, C-peptide, HOMA2 of β-cell function, HOMA2 of insulin sensitivity, HOMA of insulin resistance, and body weight.

RESULTS

Participants (56% women, mean age 38.3 years) provided complete insulin data in 131 cases. Compared with placebo, semaglutide significantly reduced fasting glucose (−0.87 mmol/L [95% CI −1.15, −0.59]; P < 0.001), improved insulin sensitivity (8.60 [5.82, 13.65]; P = 0.001), and lowered insulin resistance (−0.69 [−1.00, −0.20]; P = 0.006). Mean weight loss was 9.2 kg and mediated improvements in insulin sensitivity (estimate 7.82; P = 0.01) and insulin resistance (estimate −0.75; P = 0.01). Nonsignificant trends were observed toward reduced fasting insulin (−52.3 pmol/L; P = 0.11) and C-peptide (−182.9 pmol/L; P = 0.096), with a modest, nonsignificant increase in β-cell function (8.10; P = 0.19).

CONCLUSIONS

Semaglutide significantly improved insulin sensitivity, reduced insulin resistance, lowered fasting glucose, and promoted substantial weight loss in patients with antipsychotic-induced metabolic disturbances. Weight loss partly mediated the metabolic improvements, while β-cell function remained largely unchanged. These findings support semaglutide as a potential strategy for mitigating metabolic dysfunction in this high-risk population.

Graphical Abstract

An infographic summarises a 30-week randomised placebo controlled trial of semaglutide 1.0 milligram weekly in patients with schizophrenia, prediabetes, and obesity treated with second-generation antipsychotics. The population includes 154 individuals randomised to Active N 77 or Placebo N 77, with 74 and 69 completing. Main results show fasting glucose decreases by 0.87 millimole per litre, insulin sensitivity increases by 8.60, insulin resistance decreases by 0.69, and weight decreases by 9.2 kilogram. The conclusion states reduced insulin resistance, improved insulin sensitivity, lower fasting glucose, and weight loss.

Open in a new tab

Introduction

Individuals with schizophrenia have a significantly shortened life span, with an average reduction of 15–20 years compared with the general population (1,2). The increased mortality is largely attributed to a higher prevalence of prediabetes and metabolic syndrome (3), which elevate the risk of type 2 diabetes (T2D) (4) and premature cardiovascular disease (CVD) (5). The metabolic disturbances may stem from the psychiatric condition itself (6), which negatively impacts the ability to maintain a healthy lifestyle (7). However, treatment with second-generation antipsychotics (SGAs) (8) may also play a role, as this group of drugs is documented to promote weight gain and increase appetite shortly after treatment initiation (9). Insulin resistance is a key metabolic dysfunction that underlies the pathogenesis of obesity, T2D, and hypertension, thereby contributing to an increased cardiovascular mortality (10). Insulin resistance is characterized by a diminished biological response to insulin in insulin-sensitive tissues, such as muscle, liver, and adipose tissue, leading to glucose intolerance, dyslipidemia, and abnormal protein metabolism (11). To compensate, pancreatic β-cells increase insulin secretion, often resulting in hyperinsulinemia, which is a hallmark of insulin resistance (12).

Over the past decade, various weight loss interventions have been investigated in schizophrenia, including metformin, topiramate, and glucagon-like peptide 1 receptor agonists (GLP-1RAs), with the latter demonstrating the most promising results (13,14). The initial GLP-1RA–based trials in patients with schizophrenia used either liraglutide or exenatide (15,16), but today, semaglutide has largely replaced the early GLP-1RAs due to its superior efficacy in promoting weight loss in individuals with obesity. Additionally, semaglutide has been shown to improve cardiometabolic risk factors and reduce cardiovascular events in individuals with obesity, without diabetes, and with preexisting CVD (17) and to cause significant weight loss and favorable changes in body composition in individuals with schizophrenia treated with semaglutide 2.0 mg weekly for 36 weeks (18).

We recently performed a randomized, 30-week clinical trial that compared up to 1.0 mg weekly of semaglutide with placebo in patients with schizophrenia, overweight/obesity (BMI ≥27 kg/m2), and prediabetes (HbA1c 39–47 mmol/L) who were on stable treatment with SGAs (19). Semaglutide treatment was well tolerated and improved patients’ physical well-being without worsening their mental disorder. In addition, semaglutide reduced HbA1c by 5.07 mmol/mol, reduced body weight by 9.21 kg, and normalized HbA1c to <39 mmol/mol in 81% of semaglutide-treated patients vs. 19% in placebo-treated patients. The objective of this prespecified secondary end point study was to describe in more detail the glucose-metabolic changes in our cohort, focusing on HOMA-derived measures of β-cell function (HOMA-B) and insulin sensitivity (HOMA2-S) and to compare these changes with changes in body weight.

Research Design and Methods

This trial was approved by the Regional Committees on Health Research Ethics for Southern Denmark, Odense, Denmark (S-20200182), and conducted in accordance with the Declaration of Helsinki. The study was initiated prior to the introduction of Wegovy (semaglutide up to 2.4 mg weekly; Novo Nordisk A/S, Bagsværd, Denmark); therefore, semaglutide was tested up to doses of 1 mg weekly only. As semaglutide 1 mg weekly is not approved for the treatment of patients with schizophrenia, overweight/obesity, and prediabetes, the study protocol received prior approval from the Danish Medicines Agency (2020110216). The trial was subsequently subject to close monitoring by local Good Clinical Practice units.

Design

This investigator-initiated, double-blind, parallel-group, superiority, multicenter, randomized controlled trial (RCT) took place from January 2022 (first patient, first visit) to May 2024 (last patient, last visit) in the Region of Southern Denmark and the Region of Zealand, covering all community psychiatry centers. The majority of participants’ blood samples were collected in their homes through Project HISTORI (Home-Based Intervention With Semaglutide Treatment of Neuroleptic-Related Prediabetes).

This study reports a prespecified secondary end point. A detailed description of the study protocol has been published previously, and additional information about both the primary and secondary end points is available at ClinicalTrials.gov (identifier: NCT05193578) and in the published protocol (20). The prespecified primary end point was change in HbA1c. Other prespecified secondary end points included changes in weight, BMI, cardiovascular autonomic neuropathy, schizophrenia symptom severity, quality of life, physical activity levels, and adherence to antipsychotic treatment.

Participants

The study enrolled patients clinically diagnosed with either schizophrenia, schizotypal disorder, or schizoaffective disorder according to ICD-10 codes F20.x, F21.x, and F25.x. All patients were between 18 and 60 years of age and received SGA treatment. Furthermore, all patients had prediabetes (defined as an HbA1c of 39–47 mmol/mol) and overweight/obesity (defined by a BMI ≥27 kg/m2). A comprehensive list of all relevant inclusion and exclusion criteria can be found in Supplementary Table 1. None of the participants in the study received any compensation, aside from mileage allowances, for participation.

Procedures

Participants received semaglutide (1.34 mg/mL) up to 1.0 mg/week or a similar volume of placebo for 30 weeks with injection once weekly. Semaglutide and placebo were uptitrated according to recommendations by the manufacturer (Novo Nordisk). Novo Nordisk provided the study drug (active comparator and placebo) and randomization list but was not involved in the study.

At baseline and at 30 weeks of treatment, all participants gave fasting blood samples for assessments of HbA1c, triglycerides, cholesterol (HDL and LDL), fasting blood glucose, insulin, and C-peptide. We also registered their BMI, waist circumference, and blood pressure.

The outcomes were changes in HOMA2 estimates, calculated using HOMA2 (version 2.2.3) based on fasting blood glucose and fasting insulin levels. Samples were analyzed at the Department of Biochemistry, Odense University Hospital, according to International Federation of Clinical Chemistry and Laboratory Medicine standards. All the samples were measured en bloc after finalizing the last patient, last visit. Other variables, such as HbA1c and lipid levels, were measured at the department of biochemistry at local hospitals after ensuring that they used the same assays and platforms. All patients followed their clinical psychiatric treatment plan, with individually adjusted follow-up visits at their usual clinician/treatment facility independent of the study.

Randomization and Blinding

Block randomization (blocks of four or six participants) was used. The randomization list was provided by Novo Nordisk. Participants were assigned a randomization number upon study entry using Research Electronic Data Capture, and randomization was conducted continuously without stratification. Each number corresponded to a prepackaged set of study medication or placebo for the entire study period.

All participants and study personnel were blinded; only the pharmacy had access to the randomization list. The trial drug and placebo were visually identical, medication packs were labeled with participant numbers, and unblinded pharmacy staff had no contact with blinded investigators. Unblinding occurred only after all end points were assessed, except for one participant who died of natural causes and was later revealed to have received placebo.

Variables of Interest and Interpretation

β-Cell Function

β-Cell function was estimated using the HOMA-B, which is a widely used method to estimate pancreatic β-cell function based on fasting glucose and insulin levels. This model assumes a steady-state relationship between insulin secretion and glucose levels, providing an indirect measure of β-cell activity. HOMA-B is particularly useful due to its simplicity, but it has limitations, especially in individuals with extreme glucose levels, where its accuracy may be reduced. Despite these constraints, HOMA-B remains a valuable tool for assessing β-cell function in both research and clinical settings (21).

Insulin Sensitivity

Insulin sensitivity refers to how effectively the tissues, particularly muscle, liver, and fat cells, respond to insulin and take up glucose from the blood stream. Insulin sensitivity may be estimated by HOMA-S. HOMA-S aligns with the gold standard of the euglycemic-hyperinsulinemic clamp, making the model useful in clinical settings (22).

Statistical Analysis

The power calculation was based on HbA1c changes. Assuming that 30 weeks of semaglutide would reduce HbA1c by 0.2% and an SD of 0.35%, 65 participants in each arm were required to obtain a power of 90% with a two-sided significance level of 5%. This study was not powered to detect changes in insulin-related end points.

To assess sample reproducibility and treatment group exchangeability, we conducted descriptive analyses. Categorical outcomes are reported as numbers and percentages and analyzed using Fisher exact test or χ2 test. Continuous outcomes are described as mean (SD) or median (interquartile range) and analyzed using the Wilcoxon rank sum test or unpaired t test.

HOMA outcomes were analyzed using a mixed-effects model, which was preferred over G-computation due to the absence of time-varying exposure and confounding. A likelihood ratio test was used to assess the longitudinal effect of semaglutide. Treatment effects at weeks 15 and 30 were tested via mixed-effects models with restricted maximum likelihood estimation and a t test using the Kenward-Roger method. Models included baseline covariates and time interactions, with fit assessed via normality of residuals and random effects; bootstrapping was applied if needed. The discrimination between high and low insulin sensitivity (HOMA2-S) and β-cell function (HOMA2-B) was defined by the median of HOMA2-B = 115.3 and HOMA2-S = 63.5 in a matched background population with normal glucose tolerance (23).

Mediator analysis among the exposure, treatment assignment, mediator, weight loss, and outcome followed a counterfactual framework, adjusting for baseline by evaluating changes over time and thereafter analyzed in a cross-sectional manner. Therefore, we used G-computation, in which the Q-model was chosen to be a linear regression at each time point. The statistical tests for the three effects (natural direct, indirect, and total effect) were based on a z test. The CIs for the effects were constructed such that they realigned with the z test with robust SEs. The model assumptions for the three Q-models were checked via quantile-quantile plots. For the three models, none of them were adjusted for potential confounders except the models focusing on the association between the mediator and outcome, which were adjusted for sex.

Efficacy and safety were analyzed using an intention-to-treat approach, in which participants who received at least one dose and had postbaseline assessments where analyzed according to the assigned group. Safety analysis included all randomized participants who received at least one dose. Treatment groups were unblinded after the statistical analyses.

Statistical analyses were performed using Stata 18 (StataCorp, College Station, TX). Two-sided P values and 95% CIs are reported, with an α of 0.05. No multiple-testing adjustments were made, as analyses were prespecified.

Results

We assessed 402 potential participants for eligibility, of whom 154 were randomized to 77 in each arm (Supplementary Fig. 1). A total of 141 participants (91.5%) completed the trial (Supplementary Fig. 1), including 74 (96%) randomized to semaglutide and 67 randomized to placebo (87%). The randomization resulted in two comparable groups (Table 1). The original publication provides additional information on the study groups (19).

Table 1.

Baseline characteristics of study participants

Characteristic Active (n = 77) Placebo (n = 77) P
Age, years, mean (SD) 39.1 (10.9) 37.6 (10.6) 0.40
Female 35 (45) 52 (68) 0.004
Smoking 48 (62) 46 (61) 0.66
Alcohol >14 units weekly 1 (1) 0 (0) 1.00
Schizophrenia 77 (100) 76 (100)* 1.00
Treatment      
 Quetiapine 34 (44) 36 (47) 0.69
 Olanzapine 13 (17) 9 (12) 0.37
 Risperidone 2 (3) 8 (11) 0.047
 Ziprasidone 3 (4) 3 (4) 0.99
 Paliperidone 5 (6) 4 (5) 0.75
 Aripiprazole 23 (30) 33 (43) 0.082
 Clozapine 22 (29) 19 (25) 0.62
 Other 8 (10) 13 (17) 0.23
  1 drug 46 (60) 33 (43)  
  2 drugs 29 (38) 37 (49) 0.078
  ≥3 drugs 2 (3) 6 (8)  
Clinical characteristics, mean (SD)      
 Body weight, kg 122.28 (25.92) 119.66 (27.68) 0.55
 Waist circumference, cm 120.78 (15.54) 116.86 (14.22) 0.11
 BMI, kg/m² 40.66 (8.14) 40.16 (8.47) 0.71
 Diastolic blood pressure, mmHg 85.17 (7.74) 82.69 (8.61) 0.065
 Systolic blood pressure, mmHg 130.04 (12.98) 129.66 (12.55) 0.86
 Glucose metabolism      
  HbA1c, mmol/mol 40.94 (2.65) 41.5 (2.40) 0.17
  Fasting blood glucose, mmol/L 6.28 (0.94) 6.13 (0.95) 0.34
  Fasting C-peptide, pmol/L 1,635.6 (592.24) 1,491.93 (480.23) 0.23
  Fasting insulin, pmol/L 224.90 (133.93) 201.72 (123.52) 0.42
Open in a new tab

Data are n (%) unless otherwise indicated.

*One participant died of natural causes. Some data are reprinted from our initial study (19).

From the 141 participants completing the study, we had complete data at both time points for 131 (69 in the semaglutide group and 62 in the placebo group). Per protocol, all participants had prediabetes according to HbA1c.

Baseline levels of fasting C-peptide, fasting insulin, and fasting glucose were comparable between the groups (Table 1). Compared with placebo, semaglutide tended to reduce C-peptide (−182.93 pmol/L [95% CI, −260.36, 32.19]; P = 0.096) and tended to improve insulin (−52.30 pmol/L [−74.40, 11.61]; P = 0.11), whereas fasting plasma glucose was significantly lower (−0.87 mmol/L [−1.15, 0.59]; P < 0.001) (Table 2).

Table 2.

Main effects on outcomes

Outcome Semaglutide (active comparator) Placebo Treatment effect/difference P
Body weight, kg −9.04 (−10.75, −7.34) 0.17 (−1.65, 1.99) −9.21 (−11.68, −6.75) <0.001
Waist circumference, cm −5.58 (−5.21, −7.70) −0.61 (−1.75, 2.97) −6.19 (−9.26, −3.12) <0.001
BMI, kg/m² −2.98 (−3.53, −2.42) −0.12 (−0.72, 0.48) −2.86 (−3.71, −2.01) <0.001
Diastolic BP, mmHg −2.91 (−5.75, −0.08) −0.66 (−3.7, 2.38) −2.26 (−6.42, 1.91) 0.289
Systolic BP, mmHg −1.61 (−3.57, 0.35) −0.23 (−3.24, 2.79) −1.38 (−5.03, 2.26) 0.457
HbA1c, mmol/mol −4.11 (−4.74, −3.48) 0.96 (0.15, 1.77) −5.07 (−6.05, −4.09) <0.001
FBG, mmol/L −0.69 (−0.85, 0.54) 0.18 (0.04, 0.39) −0.87 (−1.15, 0.59) <0.001
C-peptide, pmol/L −98.75 (−260.37, 62.86) 84.19 (−91.14, 259.49) −182.93 (−260.37, 32.19) 0.096
Insulin, pmol/L −24.84 (−74.40, 24.72) 27.46 (−25.39, 80.31) −52.30 (−74.40, 11.61) 0.109
Insulin sensitivity 10.08 (5.82, 14.33) 1.48 (−2.79, 5.75) 8.60 (5.82, 13.65) 0.001
β-Cell function 8.00 (−1.66, 17.65) −0.10 (−8.54, 8.33) 8.10 (−1.66, 20.22) 0.190
Insulin resistance −0.68 (−1.00, −0.35) 0.01 (−0.44, 0.46) −0.686 (−1.00, −0.20) 0.006
Open in a new tab

Data are mean (95% CI). Boldface indicates significance at P < 0.05. BP, blood pressure; FBG, fasting blood glucose. Some data are reprinted from our initial study (19).

At baseline, HOMA estimates showed comparable β-cell function and insulin sensitivity in the two groups (Fig. 1). After 30 weeks, β-cell function had increased numerically in the semaglutide group (8.00 [95% CI, −1.66, 17.65]) and remained unchanged in the placebo group (−0.10 [−8.54, 8.33]), but overall, semaglutide caused a nonsignificant increase in β-cell function (8.10 [−1.66, 20.22]; P = 0.19) (Table 2). In contrast, insulin sensitivity increased significantly in the semaglutide-treated group (10.08 [5.82, 14.33]) compared with the placebo-treated group (1.48 [−2.79, 5.75]), resulting in a significant treatment effect averaging 8.60 (5.82, 13.65; P = 0.001) (Fig. 1).

Figure 1.

Two line charts labelled A and B show insulin sensitivity and beta cell function using H O M A percent S and H O M A percent B at Baseline and week 30. In A, the Active group rises from about 35 at Baseline to about 45 at week 30, while the Placebo group rises slightly from about 35 to about 37. In B, the Active group rises from about 157 at Baseline to about 165 at week 30, while the Placebo group remains near 157. Error bars appear at each point.

Open in a new tab

Margin plots of HOMA2-B and HOMA2-S. A and B: HOMA2 estimates of insulin sensitivity (A) and β-cell function (B) at baseline and week 30 for participants receiving semaglutide or placebo. At week 30, semaglutide significantly increased insulin sensitivity (mean difference 8.60 [95% CI 5.82, 13.65]; P = 0.001). β-Cell function showed a modest, nonsignificant improvement (mean difference 8.60 [–2.53, 19.73]; P = 0.13). Data are mean ± SE.

To contextualize our findings, we compared our study cohort with previously published data from our institution on individuals with normal glucose tolerance (23), illustrating insulin sensitivity and β-cell function using a two-dimensional distribution (Fig. 2). The majority of participants showed increased β-cell function and reduced insulin sensitivity, indicating an elevated insulin resistance.

Figure 2.

A scatter plot shows H O M A 2 percent B on the vertical axis and H O M A 2 percent S on the horizontal axis. The data points cluster at lower percent S values between about 10 and 50, with percent B ranging from about 80 to 300. Fewer points appear above 60 percent S, where percent B is mostly below 120. A vertical line marks about 63.5 on percent S, and a horizontal line marks about 115.3 on percent B.

Open in a new tab

Baseline association between HOMA-S and HOMA-B. Plot of insulin sensitivity and β-cell function of patients at baseline. Reference lines represent the median values of HOMA2-S and HOMA2-B in the background population. Red circles indicate insulinopenic phenotype; dark blue circles, classical phenotype; light blue circles, hyperinsulinemic phenotype.

Weight Loss and HOMA Values

Mediator analysis indicated that weight loss significantly mediated the effect of semaglutide on improving insulin sensitivity (estimate 7.82 [95% CI 1.83, 13.81]; P = 0.01). In contrast, weight loss was not a significant mediator for changes in fasting C-peptide (estimate −190.47 [−440.75, 59.81]; P = 0.14), changes in fasting insulin (estimate −61.13 [−146.30, 24.03]; P = 0.16), or the increase in β-cell function (estimate 9.31 [−5.21, 23.82]; P = 0.21) (Supplementary Table 2).

Antipsychotic Treatment and HOMA Values

There were no associations between the type of antipsychotic treatment and pretreatment HOMA estimates of β-cell function or insulin sensitivity (Supplementary Tables 3 and 4 and Supplementary Fig. 2).

Conclusions

In this article, we describe changes in β-cell function and insulin sensitivity in individuals with SGA-treated schizophrenia, prediabetes, and overweight/obesity. These data are based on our recently performed double-blind, placebo-controlled RCT, in which we compared 30 weeks of semaglutide up to 1.0 mg weekly versus placebo. We could demonstrate that semaglutide was well-tolerated, reduced mean body weight by 9.2 kg (∼7.5% of the body weight), and reduced HbA1c to 5.07 mmol/mol, alongside an improvement in physical quality of life, compared with placebo (19). We were also able to demonstrate that the semaglutide-treated group, alongside the reduction in HbA1c, showed improvements in fasting blood glucose, while its effects on fasting C-peptide and insulin levels did not reach statistical significance. Here, we describe changes in glucose homeostasis in further detail by demonstrating that insulin sensitivity/resistance improved significantly with semaglutide treatment, whereas β-cell function remained unchanged, and that the semaglutide-induced weight loss was the mediator of the beneficial changes in insulin sensitivity/resistance.

Hyperinsulinemia, Insulin Sensitivity, and β-Cell Function

Using HOMA2 estimates, we demonstrated that at baseline, our participants had reduced insulin sensitivity and elevated β-cell function compared with individuals without diabetes and schizophrenia (Fig. 2). Even though the participants in our trial had prediabetes and not overt T2D, the majority exhibited insulin levels similar to individuals with T2D and highest levels of insulin resistance (23). Furthermore, our data align well with a meta-analysis performed by Pillinger et al. (6), who showed that individuals with schizophrenia exhibit elevated fasting plasma glucose levels, diminished glucose tolerance, increased fasting plasma insulin levels, and enhanced insulin resistance at the onset of illness.

After 30 weeks, semaglutide treatment could be linked to a significant increase in insulin sensitivity and a corresponding decrease in insulin resistance. This finding suggests that the semaglutide-induced improvement in glucose regulation is driven, at least in part, by enhanced insulin action rather than direct effects on β-cell function.

Despite improvements in glucose metabolism (HbA1c and fasting glucose), as well as insulin sensitivity, β-cell function, as assessed by HOMA-B, remained unchanged following semaglutide treatment. While GLP-1RAs have been shown to enhance β-cell function in some studies (24,25), the current findings suggest that in our population, the primary driver of metabolic improvements was weight loss rather than direct β-cell effects.

The mediator analyses suggest that the beneficial effects of semaglutide on glucose metabolism and insulin homeostasis are largely mediated by semaglutide’s weight-reducing properties rather than direct pharmacological effects on insulin secretion or β-cell function. Thus, our findings align with evidence derived from the Semaglutide Treatment Effect in People With Obesity (STEP) and Semaglutide Effects on Cardiovascular Outcomes in People With Overweight and Obesity (SELECT) programs and with mediation analyses showing that GLP-1RA–associated improvements in glycemia, lipid parameters, and CVD risk are mediated by weight loss (26). It is important to note that semaglutide 1.0 mg weekly for 30 weeks was used in this trial. Studies using semaglutide 2.4 mg have shown an average weight loss of 15% following 68 weeks treatment (27). Using semaglutide 2.4 mg over a longer duration might have had a greater effect on glucose metabolism and insulin homeostasis.

Insulin resistance has been linked to an increased risk for CVD (28,29). As weight loss appeared to mediate the observed beneficial effects of semaglutide on insulin sensitivity/resistance, our findings suggest that the treatment of patients with schizoaffective disorder should include weight loss to prevent CVD (30).

Antipsychotic Treatment and Metabolic Outcomes

Antipsychotic medications, which are fundamental in the treatment of severe mental illnesses such as schizophrenia spectrum disorders, substantially elevate the risk of metabolic complications. These adverse effects include weight gain, dyslipidemia, and insulin resistance (3,31). As evidenced by an extensive Danish cohort study, all types of antipsychotics increase the risk of T2D, regardless of the specific agent or class, whether first or second generation (8). This notion aligns with our data, as we did not detect any correlation between specific antipsychotic drugs and HOMA values. Although treatment with clozapine and olanzapine has been associated with increases in plasma glucose (32), we did not observe any specific correlation between treatment with these drugs and the HOMA estimates. However, we acknowledge that our sample size was small and that only 22 of 154 participants received olanzapine and only 41 of 154 received clozapine. Thus, the comparison of different SGAs remains of an explorative nature, as the study was not powered to evaluate antipsychotic effects on insulin sensitivity or β-cell function.

Metabolic syndrome in patients with schizophrenia is partly driven by antipsychotics, causing increased appetite, weight gain, and consequently, insulin resistance and abdominal obesity (33). Insulin resistance can lead to T2D, dyslipidemia, and essential hypertension (34), which are key contributors to atherosclerosis. Consequently, the increased CVD mortality in patients with schizophrenia may be explained by the increased prevalence and severity of insulin resistance (5). The most effective way to treat insulin resistance is through weight loss, which reduces abdominal fat (35). Caloric restriction and increased physical activity are crucial, and there is evidence that they can reverse insulin resistance (36), but treatment with weight loss agents such as semaglutide should be considered in patients with schizophrenia and a high risk of CVD, as conventional approaches focusing on lifestyle factors, such as weight-reducing diets and physical activity, have shown limited effects (37).

Strengths and Limitations

Our study has several strengths, including its RCT design and the high rate of completion (91.5%). However, some limitations should be noted. First, our findings may not be generalizable to all patient populations, particularly those without concurrent antipsychotic treatment. Second, although we observed significant metabolic benefits, our sample size may have limited the power to detect smaller, but clinically relevant changes in β-cell function and fasting insulin levels. Third, our study duration was limited to 30 weeks, and longer follow-up may be required to determine changes in these variables. We only used semaglutide 1.0 mg weekly, and it is likely that semaglutide at doses up to 2.4 mg weekly would have shown greater effects. Furthermore, we did not use the gold standard (i.e., the hyperinsulinemic-euglycemic clamp technique) to determine the glucose metabolic condition of our participants but relied on fasting insulin and glucose levels and the corresponding HOMA estimates. However, HOMA estimates have been widely validated as a surrogate marker of insulin resistance, and this method has provided outcomes that correlate well with conventional tests of insulin resistance, such as the hyperinsulinemic-euglycemic clamp (38). Finally, our data are limited to fasting conditions, as we did not obtain stimulated measures of β-cell function, a clear limitation. Additional data on stimulated HOMA estimates would be valuable to better understand changes before and after semaglutide treatment, especially regarding GLP-1RA–mediated effects on glucose-dependent insulin secretion.

In conclusion, semaglutide treatment led to significant improvements in glycemic control and insulin sensitivity in the fasting condition, with weight loss appearing to be the primary mediator of these effects. While β-cell function did not change significantly, the observed metabolic benefits support the role of semaglutide as an effective intervention for improving metabolic health in individuals undergoing antipsychotic treatment. Given the high prevalence of insulin resistance and cardiovascular risk in schizophrenia (39), semaglutide may represent a valuable strategy for mitigating metabolic dysfunction in this at-risk population.

This article contains supplementary material online at https://doi.org/10.2337/figshare.31298458.

Article Information

Acknowledgments. The authors thank Ulla Falkner, clinical research nurse, Research Unit West (Slagelse)–Psychiatry Region Zealand; Birgitte Lundberg, clinical research nurse, Psychiatric Research Unit, Copenhagen University Hospital–Psychiatry Region Zealand; Rikke Skov Jensen, clinical research nurse, Child and Adolescent Psychiatric Research Unit, Psychiatry Region of Southern Denmark; and chemist Lise Pedersen and the Clinical Biochemistry Department, Holbæk Hospital, coordinator for blood sampling in Region Zealand During the preparation of this work, the authors used GPT-4 Turbo to correct grammar, condense text, and enhance academic writing style. Following the use of this tool, the authors formally reviewed the content for its accuracy and edited it as necessary. The authors take full responsibility for all the content of this publication.

Duality of Interest. A.A.G. reported receiving a PhD salary from Steno Diabetes Center Odense in 2021 and travel support from Novo Nordisk. N.U. reported receiving a PhD salary from Steno Diabetes Center Zealand in 2021; grant funding for the study administered by hospitals from the Novo Nordisk Foundation; salary for nurses from Slagelse research grants in 2020, 2022, and 2024; and grants from Region Zealand Health Research. S.A. reported receiving grant funding from the Novo Nordisk Foundation, Lundbeck Foundation, Region Zealand Health Research Foundation, and Slagelse Research Fund. P.G. reported receiving lecture fees paid to his institution from AstraZeneca, Bayer, Boehringer Ingelheim, and Novo Nordisk and participating on advisory boards for AstraZeneca, Bayer, Boehringer Ingelheim, Novo Nordisk, with fees paid to his institution, outside the submitted work. N.B. reported receiving grants from the Lundbeck Foundation, Novo Nordisk Foundation, and Tryg Foundation and lecture fees from Novo Nordisk and Takeda outside the submitted work. J.F. reported receiving a PhD salary from the Steno Diabetes Center Odense during the conduct of this study, nonpersonal finances from Novo Nordisk A/S as part of a scientific collaboration unrelated to the study, and speaker fees from Eli Lilly outside the submitted work. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. A.A.G. drafted the manuscript. A.A.G., N.U., S.A., P.G., N.B., and J.F. contributed to the study concept and design and obtained funding. A.A.G., N.U., N.B., and J.F. contributed to the acquisition, analysis, or interpretation of the data. N.U., S.A., P.G., N.B., and J.F. critically reviewed the manuscript for important intellectual content. S.A., P.G., N.B., and J.F. supervised the study. A.K.P. performed the statistical analysis. A.A.G. and N.U. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the Joint Congress of the European Society for Pediatric Endocrinology and European Society of Endocrinology, Copenhagen, Denmark, 10–13 May 2025.

Handling Editors. The journal editors responsible for overseeing the review of the manuscript were Cheryl A.M. Anderson and Amalia Gastaldelli.

Funding Statement

Novo Nordisk A/S provided the investigational drug and placebo. The study was funded by Region Sjælland, Steno Diabetes Center Zealand, Aase og Ejnar Danielsens Fond, Novo Nordisk Fonden grant NNF19OC0058155, and Steno Diabetes Center Odense. The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The scientists involved conducted the experiment out of general scientific interest without personal financial gain. None of the investigators involved had financial interests (including shares, direct employment, membership on advisory boards) in the drug company that produces the active drug used in the study. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

Clinical trial reg. no. NCT05193578, https://clinicaltrials.gov

Supporting information

Supplementary Material
dc252041_supp.zip (760.3KB, zip)

References

  • 1. Nordentoft M, Wahlbeck K, Hällgren J, et al. Excess mortality, causes of death and life expectancy in 270,770 patients with recent onset of mental disorders in Denmark, Finland and Sweden. PLoS One 2013;8:e55176. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Laursen TM, Munk-Olsen T, Vestergaard M. Life expectancy and cardiovascular mortality in persons with schizophrenia. Curr Opin Psychiatry 2012;25:83–88 [DOI] [PubMed] [Google Scholar]
  • 3. Mitchell AJ, Vancampfort D, Sweers K, et al. Prevalence of metabolic syndrome and metabolic abnormalities in schizophrenia and related disorders-a systematic review and meta-analysis. Schizophr Bull 2013;39:306–318 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Bushe C, Holt R. Prevalence of diabetes and impaired glucose tolerance in patients with schizophrenia. Br J Psychiatry Suppl 2004;47:s67–s71 [DOI] [PubMed] [Google Scholar]
  • 5. Kelly DL, McMahon RP, Liu F, et al. Cardiovascular disease mortality in patients with chronic schizophrenia treated with clozapine: a retrospective cohort study. J Clin Psychiatry 2010;71:304–311 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Pillinger T, Beck K, Gobjila C, et al. Impaired glucose homeostasis in first-episode schizophrenia: a systematic review and meta-analysis. JAMA Psychiatry 2017;74:261–269 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. McCreadie RG; Scottish Schizophrenia Lifestyle Group . Diet, smoking and cardiovascular risk in people with schizophrenia: descriptive study. Br J Psychiatry 2003;183:534–539 [DOI] [PubMed] [Google Scholar]
  • 8. Rajkumar AP, Horsdal HT, Wimberley T, et al. Endogenous and antipsychotic-related risks for diabetes mellitus in young people with schizophrenia: a Danish population-based cohort study. Am J Psychiatry 2017;174:686–694 [DOI] [PubMed] [Google Scholar]
  • 9. Muench J, Hamer AM. Adverse effects of antipsychotic medications. Am Fam Physician 2010;81:617–622 [PubMed] [Google Scholar]
  • 10. Johnson AMF, Olefsky JM. The origins and drivers of insulin resistance. Cell 2013;152:673–684 [DOI] [PubMed] [Google Scholar]
  • 11. Abdul-Ghani MA, DeFronzo RA. Pathogenesis of insulin resistance in skeletal muscle. J Biomed Biotechnol 2010;2010:476279. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Petersen MC, Shulman GI. Mechanisms of insulin action and insulin resistance. Physiol Rev 2018;98:2133–2223 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Zimbron J, Khandaker GM, Toschi C, et al. A systematic review and meta-analysis of randomised controlled trials of treatments for clozapine-induced obesity and metabolic syndrome. Eur Neuropsychopharmacol 2016;26:1353–1365 [DOI] [PubMed] [Google Scholar]
  • 14. Lee K, Abraham S, Cleaver R. A systematic review of licensed weight-loss medications in treating antipsychotic-induced weight gain and obesity in schizophrenia and psychosis. Gen Hosp Psychiatry 2022;78:58–67 [DOI] [PubMed] [Google Scholar]
  • 15. Larsen JR, Vedtofte L, Jakobsen MSL, et al. Effect of liraglutide treatment on prediabetes and overweight or obesity in clozapine- or olanzapine-treated patients with schizophrenia spectrum disorder: a randomized clinical trial. JAMA Psychiatry 2017;74:719–728 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Ishøy PL, Knop FK, Broberg BV, et al. Effect of GLP-1 receptor agonist treatment on body weight in obese antipsychotic-treated patients with schizophrenia: a randomized, placebo-controlled trial. Diabetes Obes Metab 2017;19:162–171 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Lincoff AM, Brown-Frandsen K, Colhoun HM, et al.; SELECT Trial Investigators . Semaglutide and cardiovascular outcomes in obesity without diabetes. N Engl J Med 2023;389:2221–2232 [DOI] [PubMed] [Google Scholar]
  • 18. Siskind D, Baker A, Arnautovska U, et al. Efficacy and safety of semaglutide versus placebo for people with schizophrenia on clozapine with obesity (COaST): a phase 2, multi-centre, participant and investigator- blinded, randomised controlled trial in Australia. Lancet Psychiatry 2025;12:493–503 [DOI] [PubMed] [Google Scholar]
  • 19. Ganeshalingam AA, Uhrenholt N, Arnfred S, et al. Semaglutide treatment of antipsychotic-treated patients with schizophrenia, prediabetes, and obesity: the HISTORI randomized clinical trial. JAMA Psychiatry 2025;82:1065–1074 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Ganeshalingam AA, Uhrenholt NG, Arnfred S, et al. Home-based Intervention with Semaglutide Treatment of Neuroleptic-Related Prediabetes (HISTORI): protocol describing a prospective, randomised, placebo controlled and double-blinded multicentre trial. BMJ Open 2024;14:e077173. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Romo-Romo A, Aguilar-Salinas CA, Gómez-Díaz RA, et al. Validity and reliability of simple surrogate indexes to evaluate beta-cell function and insulin sensitivity. Rev Med Chil 2022;150:1458–1466 [DOI] [PubMed] [Google Scholar]
  • 22. Rudvik A, Månsson M. Evaluation of surrogate measures of insulin sensitivity - correlation with gold standard is not enough. BMC Med Res Methodol 2018;18:64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Stidsen JV, Henriksen JE, Olsen MH, et al. Pathophysiology-based phenotyping in type 2 diabetes: a clinical classification tool. Diabetes Metab Res Rev 2018;34:e3005. [DOI] [PubMed] [Google Scholar]
  • 24. MacDonald PE, El-Kholy W, Riedel MJ, et al. The multiple actions of GLP-1 on the process of glucose-stimulated insulin secretion. Diabetes 2002;51(Suppl. 3):S434–S442 [DOI] [PubMed] [Google Scholar]
  • 25. Anholm C, Kumarathurai P, Pedersen LR, et al. Liraglutide effects on beta-cell, insulin sensitivity and glucose effectiveness in patients with stable coronary artery disease and newly diagnosed type 2 diabetes. Diabetes Obes Metab 2017;19:850–857 [DOI] [PubMed] [Google Scholar]
  • 26. Bergmann NC, Davies MJ, Lingvay I, et al. Semaglutide for the treatment of overweight and obesity: a review. Diabetes Obes Metab 2023;25:18–35 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Wilding JPH, Batterham RL, Calanna S, et al.; STEP 1 Study Group . Once-weekly semaglutide in adults with overweight or obesity. N Engl J Med 2021;384:989–1002 [DOI] [PubMed] [Google Scholar]
  • 28. DeFronzo RA, Abdul-Ghani M. Assessment and treatment of cardiovascular risk in prediabetes: impaired glucose tolerance and impaired fasting glucose. Am J Cardiol 2011;108:3B–24B [DOI] [PubMed] [Google Scholar]
  • 29. Wang T, Li M, Zeng T, et al. Association between insulin resistance and cardiovascular disease risk varies according to glucose tolerance status: a nationwide prospective cohort study. Diabetes Care 2022;45:1863–1872 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Horn DB, Almandoz JP, Look M, et al. What is clinically relevant weight loss for your patients and how can it be achieved? A narrative review. Postgrad Med 2022;134:359–375 [DOI] [PubMed] [Google Scholar]
  • 31. Ballon JS, Pajvani U, Freyberg Z, et al. Molecular pathophysiology of metabolic effects of antipsychotic medications. Trends Endocrinol Metab 2014;25:593–600 [DOI] [PubMed] [Google Scholar]
  • 32. Pillinger T, McCutcheon RA, Vano L, et al. 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;7:64–77 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Newcomer JW. Antipsychotic medications: metabolic and cardiovascular risk. J Clin Psychiatry 2007;68(Suppl. 4):8–13 [PubMed] [Google Scholar]
  • 34. Ventriglio A, Gentile A, Stella E, et al. Metabolic issues in patients affected by schizophrenia: clinical characteristics and medical management. Front Neurosci 2015;9:297. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Papakonstantinou E, Oikonomou C, Nychas G, et al. Effects of diet, lifestyle, chrononutrition and alternative dietary interventions on postprandial glycemia and insulin resistance. Nutrients 2022;14:823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Taylor R, Al-Mrabeh A, Sattar N. Understanding the mechanisms of reversal of type 2 diabetes. Lancet Diabetes Endocrinol 2019;7:726–736 [DOI] [PubMed] [Google Scholar]
  • 37. Speyer H, Christian Brix Nørgaard H, Birk M, et al. The CHANGE trial: no superiority of lifestyle coaching plus care coordination plus treatment as usual compared to treatment as usual alone in reducing risk of cardiovascular disease in adults with schizophrenia spectrum disorders and abdominal obesity. World Psychiatry 2016;15:155–165 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Hanley AJG, Williams K, Stern MP, et al. Homeostasis model assessment of insulin resistance in relation to the incidence of cardiovascular disease: the San Antonio heart study. Diabetes Care 2002;25:1177–1184 [DOI] [PubMed] [Google Scholar]
  • 39. Henderson DC, Vincenzi B, Andrea NV, V, et al. Pathophysiological mechanisms of increased cardiometabolic risk in people with schizophrenia and other severe mental illnesses. Lancet Psychiatry 2015;2:452–464 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material
dc252041_supp.zip (760.3KB, zip)

Articles from Diabetes Care are provided here courtesy of American Diabetes Association

RESOURCES