Quantitative computed tomography for assessing body composition in schizophrenia: a potential indicator of insulin resistance

Objective: To investigate the relationship between the Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) index and organ body composition, and to evaluate the potential of quantitative computed tomography (QCT) parameters as biomarkers for indicating insulin resistance (IR) in schizophrenia.

Materials and methods: A total of 924 patients with schizophrenia were enrolled and categorized into non-IR (HOMA-IR ≤ 2.5) and IR (HOMA-IR > 2.5) groups. Quantitative computed tomography (QCT) systematically measured parameters such as bone mineral density (BMD), hepatic fat fraction, and areas of subcutaneous adipose tissues (SAT) and visceral adipose tissues (VAT). Univariate and multivariate logistic regression analyses were performed to identify independent predictors of IR. The indicative performance of individual parameters and combined models-including a QCT parameters-only model and a model integrating QCT parameters with clinical covariates (gender, age, BMI)-was evaluated using receiver operating characteristic (ROC) curve analysis.

Results: Patients in the IR group (n=332) demonstrated significantly elevated levels of all QCT parameters compared to the non-IR group (n=592): BMD (125.27 vs. 114.01 mg/cm³), liver fat (10.14% vs. 6.61%), SAT (138.55 vs. 82.20 cm²), and VAT (222.83 vs. 131.92 cm²) (all P < 0.001). While all four parameters were significant in univariate analysis, multivariate analysis adjusted for age, sex, and BMI identified VAT (Adjusted OR: 1.013, 95% CI: 1.009–1.016, P < 0.001) and liver fat content (Adjusted OR: 1.092, 95% CI: 1.051–1.135, P < 0.001) as the only independent indicators. The area under the curve (AUC) for individual parameters was 0.790 for VAT, 0.769 for SAT, 0.664 for liver fat, and 0.574 for BMD. The combined QCT parameters-only model achieved an AUC of 0.834, which was significantly improved to 0.852 (P = 0.001 by DeLong’s test) after incorporating clinical covariates.

Conclusion: Visceral adipose tissue and liver fat content, as measured by QCT, are robust and independent indicator of insulin resistance in patients with schizophrenia. A model combining QCT parameters with basic clinical covariates offers the best predictive performance for identifying IR risk in this population, highlighting the critical role of ectopic fat deposition in metabolic dysregulation.

Introduction

Schizophrenia, a complex neuropsychiatric disorder, is increasingly emerging as a focus of research due to its association with metabolic disturbances - most notably insulin resistance (IR). This relationship is multifaceted, involving genetic, biochemical, and pharmacological factors. Numerous studies have consistently shown that IR is more prevalent in individuals with schizophrenia compared to healthy populations (1–3). IR, a critical factor in metabolic syndrome, greatly impacts both physical and mental health. Many antipsychotics commonly used for treating schizophrenia, including quetiapine and olanzapine, are associated with IR and metabolic syndrome (4, 5). A notable incidence of IR is found in first-episode drug-naïve schizophrenia patients, indicating that factors beyond medication might also play a role in IR within this group (6). ​These results suggest that metabolic dysregulation may be an intrinsic aspect of schizophrenia, potentially exacerbated by antipsychotics. Mendelian randomization studies have indicated a probable causal relationship between IR and the risk of schizophrenia (3, 7). Insulin is an essential hormone responsible for regulating glucose and energy balance in the body, especially impacting the brain, which is sensitive to insulin (8, 9). Research suggests that cognitive impairment in schizophrenia may be linked to brain glucose and insulin resistance, indicating that brain insulin resistance could play a role in memory deficits in these patients (10).

In clinical practice, the Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) is commonly used to evaluate insulin sensitivity. The HOMA-IR method assesses insulin sensitivity by dividing the product of fasting insulin and glucose by a constant. In standard clinical practice, its use is limited due to the fact that fasting plasma insulin levels are not commonly assessed (11).

Quantitative computed tomography (QCT) is an imaging technique used to evaluate body composition, including bone mineral density (BMD), liver fat content, visceral adipose tissue (VAT), and subcutaneous adipose tissue (SAT) (12–14). ​These quantitative parameters derived from QCT can be assessed simultaneously during chest, abdominal and/or lumbar exams without additional CT scans, radiation exposure, scan duration, cost, or patient inconvenience. One study suggests that VAT and SAT, measured with QCT, are predictors of metabolic syndrome (15). Evidence from several studies suggests that VAT and BMD are related to IR (12, 14, 16).

At present, the quantitative parameters of QCT for predicting IR in Chinese patients with schizophrenia have not been fully investigated. The connection between BMD, liver fat content, abdominal adiposity, and IR in schizophrenia patients is not yet fully understood. This study investigates the link between the HOMA-IR index and organ body composition, and evaluates the potential of QCT parameters as biomarkers for indicating insulin resistance in schizophrenia.

Materials and methods

Participants

Approved by the Institutional Review Board of the Shanghai Mental Health Center, this study was conducted between June 2019 and April 2020. All patients met the DSM-IV criteria for schizophrenia or schizophreniform disorder. Exclusion criteria include: 1) Patients with severe obesity-related comorbidities due to somatic or endocrine disorders; 2) Pregnant, lactating, or planning pregnancy; 3) Inpatients with alcohol or drug dependence as per ICD-10; 4) Patients with thoracolumbar spine metallic implants; 5) Patients with a history of hepatic surgery, significant hepatic lesions, or cirrhosis; 6) Patients unable to specify duration; 7) Patients on lipid-lowering therapy; 8) Patients with hyperthyroidism. In accordance with the specified criteria, a total of 924 subjects with schizophrenia (SCZ) were enrolled.

Demographic and clinical biochemical parameters

The study utilized the following variables: age (years), gender, body mass index (BMI) (kg/m²), course (years), fasting glucose (mmol/L), fasting insulin (pmol/L), triglycerides (TAG) (mmol/L), cholesterol (Cho) (mmol/L), high-density lipoprotein (HDL) (mmol/L), low-density lipoprotein (LDL) (mmol/L), apolipoprotein A (ApoA) (g/L), apolipoprotein B (ApoB) (g/L), apolipoprotein E (ApoE) (mg/L), lipoprotein A (LppA) (mg/L), T3 (nmol/L), T4 (nmol/L), FT3 (pmol/L), FT4 (pmol/L), and TSH (uIU/ml).The Homeostasis Model Assessment for Insulin Resistance (HOMA-IR) was employed to evaluate insulin resistance status. HOMA-IR quantifies insulin resistance using the formula: fasting insulin (pmol/L) × fasting glucose (mmol/L)/135 (17). A HOMA-IR value > 2.5 was defined as the cutoff threshold for identifying insulin resistance (11). Schizophrenia patients were categorized into non-insulin-resistant (non-IR) and insulin-resistant (IR) groups based on their HOMA-IR values, with thresholds of ≤ 2.5 and > 2.5, respectively.

QCT examinations and image analysis

Abdominal QCT scans were performed using a 64-slice CT scanner (uCT 760, United Imaging Healthcare, Shanghai, China) alongside a Mindways QCT calibration phantom (Mindways Software Inc., Austin, TX, USA).CT scanning parameters included a tube voltage of 120 kV, tube current of 200 mAs, gantry rotation speed of 0.5 s, pitch of 1.0875 mm, a 512 × 512 matrix, and a scanning field of view of 500 mm. CT image reconstruction utilized a standard algorithm with a slice thickness of 1.5 mm and interslice spacing of 1 mm.

The images were analyzed using the 3D spine function module of QCT Pro Version 6.1 software (Mindways Software Inc., Austin, TX, USA) after being transferred to a QCT bone density measurement workstation. Trabecular bone mineral density (BMD) in the lumbar spine (L1-L2) was measured using QCT. The mean BMD was derived from the two vertebral bodies, excluding areas with fractures, hemangiomas, or focal lytic or sclerotic lesions (Figure 1A). To evaluate liver fat content, manually delineated circular regions of interest (ROIs) were positioned on the axial liver slice, with each ROI encompassing an approximate area of 300 mm² while meticulously avoiding bile ducts, vascular structures, and hepatic parenchymal margins (Figure 1B).The mean value of these ROIs was calculated to quantify liver fat content. The amount of abdominal subcutaneous fat (SAT) and visceral fat (VAT) was measured at the L2-L3 intervertebral space level (Figure 1C).

Figure 1

Figure 1. Diagram illustrating the measurement of QCT parameters. (A) illustrates the bone mineral density (BMD) region of interest (ROI); (B) highlights the liver fat content ROIs; (C) displays visceral fat as the blue area within the green line and subcutaneous fat as the blue area outside the green line.

Statistical analysis

Data distribution normality and variance homogeneity were assessed using the Kolmogorov-Smirnov and Levene’s tests, respectively. For quantitative variables, comparisons were performed using independent samples t-tests, while categorical variables were analyzed via chi-squared tests. Variables showing statistical significance in the univariate analysis were incorporated into the multivariate logistic regression. ROC curves were generated, and metrics such as AUC, optimal cutoff values, accuracy, sensitivity, specificity, PPV, and NPV were computed. Statistical analyses were performed using R software (version 4.1.3), considering P < 0.05 as the threshold for significance.

Results

Demographic and clinical biochemical parameters

Table 1 displays demographic and clinical biochemical characteristics. Differences were observed in age, BMI, course, fasting glucose, fasting insulin, TAG, Cho, HDL, LDL, ApoA, ApoE, and T4 between the non-IR and IR groups, while no differences were found in gender, ApoB, LppA, T3, FT3, FT4, and TSH.

Table 1

Table 1. Demographic and clinical biochemical features of individuals with schizophrenia categorized by insulin resistance status.

QCT quantitative parameters

Table 2 presents the QCT quantitative parameters for both the non-IR and IR groups. The IR group demonstrated significantly higher BMD, liver fat content, SAT, and VAT than the non-IR group (Figure 2).

Table 2

Table 2. QCT quantitative parameters of individuals with schizophrenia categorized by insulin resistance status.

Figure 2

Figure 2. Three violin plots compare non-insulin resistant (non-IR) and insulin resistant (IR) groups. Plot (A) shows bone mineral density (BMD), with IR having slightly higher values. Plot (B) depicts liver fat content, showing IR with significantly higher levels. Plot (C) presents subcutaneous (SAT) and visceral adipose tissue (VAT) areas, with IR having higher values for both, especially VAT. ***P < 0.001.

Indicative capability of QCT quantitative parameters for insulin resistance status

Univariate logistic regression analysis indicated that BMD, liver fat content, SAT, and VAT were significantly associated with IR risk. The multivariate analysis (Table 3) confirmed these factors as independent risk factors for IR. However, after adjustment for age, sex, and BMI in a subsequent model, only VAT and liver fat content remained statistically significant independent indicators of IR (Table 4). SAT was no longer an independent risk factor after accounting for these covariates.

Table 3

Table 3. Logistic regression analysis of the QCT quantitative parameters to identify factors affecting insulin resistance.

Table 4

Table 4. Multivariable logistic regression analysis for insulin resistance adjusted for clinical covariates.

ROC analysis indicated that VAT achieved an AUC of 0.790 (95% CI: 0.760-0.821), SAT had an AUC of 0.769 (95% CI: 0.738-0.800), liver fat content recorded an AUC of 0.664 (95% CI: 0.625-0.703), and BMD showed an AUC of 0.574 (95% CI: 0.536-0.613) (Table 5, Figure 3A). The logistic regression model based on QCT parameters achieved a higher AUC of 0.834 (95% CI: 0.807–0.862) for indicating insulin resistance. Moreover, the model incorporating both QCT parameters and clinical covariates attained the highest AUC of 0.852 (95% CI: 0.825–0.878) (Table 5, Figure 3B). DeLong’s test indicated that the difference in AUC between the QCT parameters-only model and the QCT parameters and clinical covariates model was statistically significant (P = 0.001). This finding demonstrated that the addition of clinical covariates (gender, age, BMI) significantly improved the indicative performance of the model based on QCT parameters.

Table 5

Table 5. Predictive capability of QCT quantitative parameters for insulin resistance status.

Figure 3

Figure 3. Two ROC curve graphs comparing performance. Graph (A) shows curves for BMD, livefat, SAT, and VAT, with AUCs of 0.574, 0.664, 0.769, and 0.790, respectively. Graph (B) shows curves for QCT and QCT plus clinical covariates, with AUCs of 0.834 and 0.852, respectively. Both graphs plot sensitivity (TPR) against 1-specificity (FPR).

To validate the rationality of variable selection in the indicative models, we first examined pairwise relationships between continuous variables. The correlation matrix for the combined model (Figure 4A) showed no strong linear correlations (|r| > 0.8) among off-diagonal variables, preliminarily supporting variable independence and providing a basis for further multicollinearity assessment. We then performed variance inflation factor (VIF) analysis to systematically evaluate multicollinearity, using a VIF threshold of 5 to indicate significant collinearity. In the QCT parameters only model, all VIF values were below 5 (Figure 4B), indicating no substantial multicollinearity and supporting model stability and reliability. Similarly, in the combined model incorporating QCT parameters and clinical covariates, all variables also had VIF values below 5 (Figure 4C), confirming that adding clinical covariates did not introduce severe multicollinearity.

Figure 4

Figure 4. Panel (A) shows a correlation matrix of variables, including VAT, SAT, livefat, BMD, BMI, and age, using a color gradient to indicate correlation strength. Panels (B, C) display variance inflation factor (VIF) bar charts for the Combined QCT Model and Combined QCT and Clinical Covariates Model, charts highlight no multicollinearity concern as all VIF values are below 5.

Discussion

Our analysis demonstrated that patients with insulin resistance (IR) exhibited significantly higher bone mineral density (BMD), liver fat content, subcutaneous adipose tissue (SAT), and visceral adipose tissue (VAT) compared to those without IR (all P < 0.001). Univariate and multivariate logistic regression analyses, incorporating only QCT parameters, identified all four parameters as significant factors associated with IR. However, after adjustment for clinical covariates—gender, age, and body mass index (BMI)—only VAT and liver fat content remained statistically significant independent predictors. This finding suggests that the apparent association between SAT and IR is substantially confounded by these clinical factors, and highlights that the contribution of adipose tissue to IR in schizophrenia is primarily driven by its ectopic deposition in specific depots, particularly the visceral compartment and the liver. Furthermore, ROC curve analysis underscored the clinical utility of these measures: a combined model integrating QCT parameters with clinical covariates yielded the highest predictive accuracy for IR, with an AUC of 0.852 (95% CI: 0.825–0.878).

Insulin secreted by pancreatic β-cells is essential for glucose homeostasis (18). Insulin resistance results in elevated circulating insulin concentrations as a compensatory mechanism to maintain normoglycemia and prevent hyperglycemia. Insulin, an anabolic hormone, is crucial for maintaining and improving bone mineral density and structural integrity (19, 20). Conversely, bone tissue may modulate the body’s sensitivity to insulin (16, 20, 21). Recent research indicates that bones, beyond their structural support role, function as an endocrine organ influencing energy metabolism. Osteoblasts, responsible for bone synthesis and mineralization, possess insulin receptors that, upon activation by insulin, trigger postnatal bone acquisition and enhance osteocalcin production, which subsequently influences pancreatic insulin secretion to maintain glucose homeostasis (20). Osteoclasts may contribute to glucose uptake-associated insulin resistance through the secretion of resistin (21). The concept that interfering with bone resorption adversely affects glucose balance is crucial in the medical field, given that most osteoporosis medications target this bone remodeling process (22). Studies on various populations have explored the link between insulin resistance and BMD, but the results have been inconsistent. In non-diabetic Caucasian menopausal women, both total hip and lumbar BMD showed a significant positive correlation with HOMA-IR, which persisted after adjusting for body weight (23). In a study involving South Korean men, HOMA-IR was found to have a significant negative correlation with BMD (24). In healthy, non-obese, menopausal Chinese-Singaporean women, an inverse association was found between HOMA-IR and lumbar spine (L1–L4) BMD, even after controlling for lean body mass and age (16). Our study revealed that although patients in the IR group exhibited significantly higher BMD in univariate analysis, this association lost significance after adjusting for age, gender, and BMI. This finding provides important insight into the complex interplay between bone and energy metabolism. The loss of independent significance of BMD suggests that its apparent association with IR is likely not a direct causal relationship, but rather mediated or confounded by other factors.

Liver fat content and insulin resistance play key roles in the development of metabolic disorders like nonalcoholic fatty liver disease (NAFLD) and type 2 diabetes mellitus (T2DM). The interplay between hepatic fat content and insulin resistance is intricate, with evidence indicating that elevated hepatic lipid accumulation drives insulin resistance, which in turn exacerbates hepatic steatosis. Metformin, a well-known antidiabetic medication, has been shown to significantly reduce liver fat content in patients with schizophrenia experiencing olanzapine-induced weight gain (25). The study identified a positive correlation between liver fat changes and HOMA-IR, indicating that metformin’s influence on liver fat might be mediated by its effect on insulin resistance (25). In patients newly diagnosed with T2DM, elevated liver fat content was strongly correlated with increased HOMA-IR, highlighting a significant association between hepatic steatosis and insulin resistance. This study also suggested that liver fat content could serve as a potential indicator for the prediction of T2DM, emphasizing the importance of monitoring liver fat in at-risk populations (26). The liver-alpha cell axis, involving liver fat and glucagon interactions, correlates with HOMA-IR, highlighting the complex relationship between liver fat content and insulin resistance (27). Our results aligned with previous research, indicating hepatic fat accumulation is a key risk factor for insulin resistance in schizophrenia patients.

Research indicates that different fat compartments affect insulin sensitivity and lipid levels differently, with excess visceral adipose tissue (VAT) posing a greater risk for insulin resistance and hyperlipidemia compared to excess subcutaneous adipose tissue (SAT) (28, 29). Research indicates that adipose tissue releases mediators like TNF-α, IL-6, leptin, and adiponectin, which affect insulin function in target tissues (30). The Framingham Heart Study indicates that increased volumes of VAT and SAT negatively correlate with total adiponectin and resistin levels, while positively associating with systemic inflammation markers like IL-6 and C-Reactive protein (31, 32). Uncoupling protein 1 (UCP1) expression is notably higher in visceral adipose tissue (VAT) than in subcutaneous adipose tissue (SAT), with a negative correlation observed between UCP1 expression in SAT and HOMA-IR (33). In human white adipocytes, UCP1 mRNA expression is downregulated by pro-inflammatory stimuli while being upregulated by metabolically enhancing agents (33). The Dallas Heart study identified a positive correlation between VAT, SAT, and HOMA-IR in univariable analyses. In multivariable analyses, VAT still showed an association with insulin resistance markers, but the link with SAT was not significant after adjustment for BMI (34). Unlike the general population, the link between SAT/VAT and insulin resistance remains unexplored in individuals with schizophrenia. In our study, the loss of independent predictive value by SAT after adjustment for gender, age, and BMI is highly informative. This suggests that the risk associated with general adiposity (as captured by BMI) and subcutaneous fat may be largely mediated by its correlation with the more metabolically active visceral and hepatic fat. Therefore, our study highlights that VAT and liver fat, as measured by QCT, are the core adiposity-related drivers of insulin resistance in this patient population, providing distinct information beyond BMI.

Despite its strengths, this study is not without limitations. First, its cross-sectional and single-center design precludes establishing a causal relationship between body composition and insulin resistance, thereby necessitating validation in longitudinal studies. Second, the absence of a healthy control group impedes our ability to determine whether the roles of VAT and liver fat are specific to schizophrenia or reflect a general metabolic pathway. Furthermore, we were unable to fully adjust for key confounding factors, including specific antipsychotic treatment regimens, lifestyle variables (e.g., dietary patterns and physical activity), and inflammatory biomarkers. To confirm our findings and clarify the underlying mechanisms, future prospective studies—incorporating healthy control groups and rigorously collecting these clinical and biochemical data—are therefore warranted.

In conclusion, among the QCT parameters evaluated, visceral adipose tissue and liver fat content were identified as the most robust and independent predictors of insulin resistance in patients with schizophrenia, even after accounting for age, sex, and BMI. This underscores the critical importance of ectopic fat deposition in the pathogenesis of metabolic disturbances in this vulnerable population.

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement

The studies involving humans were approved by the Institutional Review Board of the Shanghai Mental Health Center. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

XK: Conceptualization, Formal Analysis, Funding acquisition, Methodology, Software, Visualization, Writing – original draft, Writing – review & editing. FZ: Data curation, Writing – original draft. MH: Data curation, Writing – original draft. JZ: Data curation, Writing – original draft. CD: Data curation, Writing – original draft. LY: Data curation, Writing – original draft. QY: Data curation, Writing – original draft. LC: Data curation, Writing – original draft. ZX: Data curation, Writing – original draft. XW: Data curation, Writing – original draft. JW: Conceptualization, Project administration, Resources, Supervision, Validation, Writing – original draft.

Funding

The author(s) declare financial support was received for the research and/or publication of this article. This research received support from the China Postdoctoral Science Foundation (2021M702170).

Acknowledgments

We express our gratitude to all participants involved in this study. The authors have stated that no conflicts of interest exist concerning this study.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that no Generative AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

1. Pillinger T, Beck K, Gobjila C, Donocik JG, Jauhar S, and Howes OD. Impaired glucose homeostasis in first-episode schizophrenia: A systematic review and meta-analysis. JAMA Psychiatry. (2017) 74:261–9. doi: 10.1001/jamapsychiatry.2016.3803, PMID: 28097367

PubMed Abstract | Crossref Full Text | Google Scholar

2. Lin C, Chen K, Zhang R, Fu W, Yu J, Gao L, et al. The prevalence, risk factors, and clinical characteristics of insulin resistance in Chinese patients with schizophrenia. Compr Psychiatry. (2020) 96:152145. doi: 10.1016/j.comppsych.2019.152145, PMID: 31710880

PubMed Abstract | Crossref Full Text | Google Scholar

3. Zhuo C, Zhang Q, Wang L, Ma X, Li R, Ping J, et al. Insulin resistance/diabetes and schizophrenia: potential shared genetic factors and implications for better management of patients with schizophrenia. CNS Drugs. (2024) 38:33–44. doi: 10.1007/s40263-023-01057-w, PMID: 38097908

PubMed Abstract | Crossref Full Text | Google Scholar

4. Correll CU, Robinson DG, Schooler NR, Brunette MF, Mueser KT, Rosenheck RA, et al. Cardiometabolic risk in patients with first-episode schizophrenia spectrum disorders: baseline results from the RAISE-ETP study. JAMA Psychiatry. (2014) 71:1350–63. doi: 10.1001/jamapsychiatry.2014.1314, PMID: 25321337

PubMed Abstract | Crossref Full Text | Google Scholar

5. Yu L, Wu S, Deng Y, Lei J, Yu L, and Li W. Insulin resistance induced by olanzapine and other second-generation antipsychotics in Chinese patients with schizophrenia: a comparative review and meta-analysis. Eur J Clin Pharmacol. (2019) 75:1621–9. doi: 10.1007/s00228-019-02739-5, PMID: 31428814

PubMed Abstract | Crossref Full Text | Google Scholar

6. Petrikis P, Tigas S, Tzallas AT, Papadopoulos I, Skapinakis P, and Mavreas V. Parameters of glucose and lipid metabolism at the fasted state in drug-naïve first-episode patients with psychosis: Evidence for insulin resistance. Psychiatry Res. (2015) 229:901–4. doi: 10.1016/j.psychres.2015.07.041, PMID: 26279127

PubMed Abstract | Crossref Full Text | Google Scholar

7. Li Z, Chen P, Chen J, Xu Y, Wang Q, Li X, et al. Glucose and insulin-related traits, type 2 diabetes and risk of schizophrenia: A mendelian randomization study. EBioMedicine. (2018) 34:182–8. doi: 10.1016/j.ebiom.2018.07.037, PMID: 30100396

PubMed Abstract | Crossref Full Text | Google Scholar

8. Kullmann S, Heni M, Hallschmid M, Fritsche A, Preissl H, and Häring HU. Brain insulin resistance at the crossroads of metabolic and cognitive disorders in humans. Physiol Rev. (2016) 96:1169–209. doi: 10.1152/physrev.00032.2015, PMID: 27489306

PubMed Abstract | Crossref Full Text | Google Scholar

9. Cetinkalp S, Simsir IY, and Ertek S. Insulin resistance in brain and possible therapeutic approaches. Curr Vasc Pharmacol. (2014) 12:553–64. doi: 10.2174/1570161112999140206130426, PMID: 23627981

PubMed Abstract | Crossref Full Text | Google Scholar

10. Wijtenburg SA, Kapogiannis D, Korenic SA, Mullins RJ, Tran J, Gaston FE, et al. Brain insulin resistance and altered brain glucose are related to memory impairments in schizophrenia. Schizophr Res. (2019) 208:324–30. doi: 10.1016/j.schres.2019.01.031, PMID: 30760413

PubMed Abstract | Crossref Full Text | Google Scholar

11. Tsai SF, Yang CT, Liu WJ, and Lee CL. Development and validation of an insulin resistance model for a population without diabetes mellitus and its clinical implication: a prospective cohort study. EClinicalMedicine. (2023) 58:101934. doi: 10.1016/j.eclinm.2023.101934, PMID: 37090441

PubMed Abstract | Crossref Full Text | Google Scholar

12. Liu Y, Gao L, Wu M, Yang B, Ren D, Zhang Z, et al. Effect of adipose tissue deposition on insulin resistance in middle-aged and elderly women: Based on QCT and MRI mDIXON-Quant. J Diabetes Invest. (2025) 16:292–7. doi: 10.1111/jdi.14352, PMID: 39540298

PubMed Abstract | Crossref Full Text | Google Scholar

13. Li YX, Sang YQ, Sun Y, Liu XK, Geng HF, Zha M, et al. Pancreatic Fat is not significantly correlated with β-cell Dysfunction in Patients with new-onset Type 2 Diabetes Mellitus using quantitative Computed Tomography. Int J Med Sci. (2020) 17:1673–82. doi: 10.7150/ijms.46395, PMID: 32714070

PubMed Abstract | Crossref Full Text | Google Scholar

14. Li Z, Yue Q, Chen Y, Zhang Y, and Wan Y. Quantitative assessment of pancreatic fat by quantitative CT in type 2 diabetes mellitus. Int J Gen Med. (2022) 15:7977–84. doi: 10.2147/ijgm.S383783, PMID: 36337672

PubMed Abstract | Crossref Full Text | Google Scholar

15. Li C, Xu B, Chen M, and Zhang Y. Evaluation for performance of body composition index based on quantitative computed tomography in the prediction of metabolic syndrome. Metab Syndr Relate Disord. (2024) 22:287–94. doi: 10.1089/met.2023.0265, PMID: 38452164

PubMed Abstract | Crossref Full Text | Google Scholar

16. Kalimeri M, Leek F, Wang NX, Koh HR, Roy NC, Cameron-Smith D, et al. Association of insulin resistance with bone strength and bone turnover in menopausal chinese-Singaporean women without diabetes. Int J Environ Res Public Health. (2018) 15:889. doi: 10.3390/ijerph15050889, PMID: 29710852

PubMed Abstract | Crossref Full Text | Google Scholar

17. Knopp JL, Holder-Pearson L, and Chase JG. Insulin units and conversion factors: A story of truth, boots, and faster half-truths. J Diabetes Sci Technol. (2019) 13:597–600. doi: 10.1177/1932296818805074, PMID: 30318910

PubMed Abstract | Crossref Full Text | Google Scholar

18. Stuhlmann T, Planells-Cases R, and Jentsch TJ. LRRC8/VRAC anion channels enhance β-cell glucose sensing and insulin secretion. Nat Commun. (2018) 9:1974. doi: 10.1038/s41467-018-04353-y, PMID: 29773801

PubMed Abstract | Crossref Full Text | Google Scholar

19. Thrailkill KM, Lumpkin CK Jr., Bunn RC, Kemp SF, and Fowlkes JL. Is insulin an anabolic agent in bone? Dissecting the diabetic bone for clues. Am J Physiol Endocrinol Metab. (2005) 289:E735–45. doi: 10.1152/ajpendo.00159.2005, PMID: 16215165

PubMed Abstract | Crossref Full Text | Google Scholar

20. Fulzele K, Riddle RC, DiGirolamo DJ, Cao X, Wan C, Chen D, et al. Insulin receptor signaling in osteoblasts regulates postnatal bone acquisition and body composition. Cell. (2010) 142:309–19. doi: 10.1016/j.cell.2010.06.002, PMID: 20655471

PubMed Abstract | Crossref Full Text | Google Scholar

21. Li X, Sun F, Lu J, Zhang J, Wang J, Zhu H, et al. Osteoclasts may affect glucose uptake-related insulin resistance by secreting resistin. Diabet Metab Syndr Obes: Targets Ther. (2021) 14:3461–70. doi: 10.2147/dmso.S316964, PMID: 34366677

PubMed Abstract | Crossref Full Text | Google Scholar

22. Ferron M, Wei J, Yoshizawa T, Del Fattore A, DePinho RA, Teti A, et al. Insulin signaling in osteoblasts integrates bone remodeling and energy metabolism. Cell. (2010) 142:296–308. doi: 10.1016/j.cell.2010.06.003, PMID: 20655470

PubMed Abstract | Crossref Full Text | Google Scholar

23. Shanbhogue VV, Finkelstein JS, Bouxsein ML, and Yu EW. Association between insulin resistance and bone structure in nondiabetic postmenopausal women. J Clin Endocrinol Metab. (2016) 101:3114–22. doi: 10.1210/jc.2016-1726, PMID: 27243136

PubMed Abstract | Crossref Full Text | Google Scholar

24. Shin D, Kim S, Kim KH, Lee K, and Park SM. Association between insulin resistance and bone mass in men. J Clin Endocrinol Metab. (2014) 99:988–95. doi: 10.1210/jc.2013-3338, PMID: 24423302

PubMed Abstract | Crossref Full Text | Google Scholar

25. Wang L, Chen Y, Sui YC, Tan XQ, Zhou Z, Li N, et al. Metformin attenuates liver fat content: finding from schizophrenia patients with olanzapine-induced weight gain. Clin Psychopharmacol Neurosci: Off Sci J Kor Coll Neuropsychopharmacol. (2020) 18:67–74. doi: 10.9758/cpn.2020.18.1.67, PMID: 31958907

PubMed Abstract | Crossref Full Text | Google Scholar

26. Wang YS, Ye J, Cao YH, Zhang R, Han XF, Zou LL, et al. Association of [(1)H]-MRS quantified liver fat content with glucose metabolism status. Diabetol Metab Syndr. (2020) 12:51. doi: 10.1186/s13098-020-00558-8, PMID: 32528557

PubMed Abstract | Crossref Full Text | Google Scholar

27. Gar C, Haschka SJ, Kern-Matschilles S, Rauch B, Sacco V, Prehn C, et al. The liver-alpha cell axis associates with liver fat and insulin resistance: a validation study in women with non-steatotic liver fat levels. Diabetologia. (2021) 64:512–20. doi: 10.1007/s00125-020-05334-x, PMID: 33275161

PubMed Abstract | Crossref Full Text | Google Scholar

28. Dhaliwal R, Shepherd JA, El Ghormli L, Copeland KC, Geffner ME, Higgins J, et al. Changes in visceral and subcutaneous fat in youth with type 2 diabetes in the TODAY study. Diabetes Care. (2019) 42:1549–59. doi: 10.2337/dc18-1935, PMID: 31167889

PubMed Abstract | Crossref Full Text | Google Scholar

29. Bódis K and Roden M. Energy metabolism of white adipose tissue and insulin resistance in humans. Eur J Clin Invest. (2018) 48:e13017. doi: 10.1111/eci.13017, PMID: 30107041

PubMed Abstract | Crossref Full Text | Google Scholar

30. Dang SY, Leng Y, Wang ZX, Xiao X, Zhang X, Wen T, et al. Exosomal transfer of obesity adipose tissue for decreased miR-141-3p mediate insulin resistance of hepatocytes. Int J Biol Sci. (2019) 15:351–68. doi: 10.7150/ijbs.28522, PMID: 30745826

PubMed Abstract | Crossref Full Text | Google Scholar

31. Pou KM, Massaro JM, Hoffmann U, Vasan RS, Maurovich-Horvat P, Larson MG, et al. Visceral and subcutaneous adipose tissue volumes are cross-sectionally related to markers of inflammation and oxidative stress: the Framingham Heart Study. Circulation. (2007) 116:1234–41. doi: 10.1161/circulationaha.107.710509, PMID: 17709633

PubMed Abstract | Crossref Full Text | Google Scholar

32. Jain SH, Massaro JM, Hoffmann U, Rosito GA, Vasan RS, Raji A, et al. Cross-sectional associations between abdominal and thoracic adipose tissue compartments and adiponectin and resistin in the Framingham Heart Study. Diabetes Care. (2009) 32:903–8. doi: 10.2337/dc08-1733, PMID: 19223612

PubMed Abstract | Crossref Full Text | Google Scholar

33. Lim J, Park HS, Kim J, Jang YJ, Kim JH, Lee Y, et al. Depot-specific UCP1 expression in human white adipose tissue and its association with obesity-related markers. Int J Obes. (2020) 44:697–70. doi: 10.1038/s41366-020-0528-4, PMID: 31965068

PubMed Abstract | Crossref Full Text | Google Scholar

34. Neeland IJ, Ayers CR, Rohatgi AK, Turer AT, Berry JD, Das SR, et al. Associations of visceral and abdominal subcutaneous adipose tissue with markers of cardiac and metabolic risk in obese adults. Obes (Silver Spring Md). (2013) 21:E439–47. doi: 10.1002/oby.20135, PMID: 23687099

PubMed Abstract | Crossref Full Text | Google Scholar