Combined aerobic and resistance exercise improves glucose metabolism and insulin resistance in children and adolescents with overweight or obesity: a meta-analysis Authors: Zhihao Qi, Xianyang Xin, Qinzheng Liu, Xiangyun Xu, Yongqing Guo Published: 2026 ABSTRACT AbstractChildren and adolescents with overweight or obesity commonly exhibit impaired glucose metabolism and insulin resistance, increasing their future metabolic risk. Although combined aerobic and resistance exercise is a promising non-pharmacological strategy for pediatric obesity management, its effects on glucose metabolism and insulin resistance remain inconclusive. This systematic review and meta-analysis evaluated randomized controlled trials of combined aerobic and resistance exercise in children and adolescents aged 18 years or younger with overweight or obesity. PubMed, Cochrane Library, Web of Science, and Embase were searched from inception to March 1, 2026. Risk of bias was assessed using RoB 2, and certainty of evidence using GRADE. Eighteen trials involving 682 participants were included. Compared with control conditions, combined aerobic and resistance exercise significantly reduced fasting plasma glucose and fasting insulin, and may be associated with reductions in HOMA-IR. Subgroup and meta-regression analyses suggested that effects varied by sex, session duration, age, and continent. Evidence certainty was low for fasting plasma glucose and fasting insulin and very low for HOMA-IR. Combined aerobic and resistance exercise may improve glucose metabolism and alleviate insulin resistance in this population, although the findings for HOMA-IR should be interpreted cautiously because of the very low certainty of evidence.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1038/s41598-026-59248-6.Keywords: Children and adolescents, Obesity, Combined aerobic and resistance exercise, Glucose metabolism, Insulin resistance, Meta-analysisSubject terms: Diseases, Endocrinology, Health care, Medical research Introduction Improvements in living standards and changes in dietary habits have made overweight and obesity in children a major global health challenge in the 21st century1,2. Between 1975 and 2016, the global prevalence of obesity among children and adolescents aged 5–19 years increased from 0.7% to 5.6% in girls and from 0.9% to 7.8% in boys3. The World Obesity Federation has projected that 206 million children and adolescents aged 5–19 years will be living with obesity by 2025, and that this number will rise to 254 million by 20304. Obesity in childhood and adolescence has been identified as a trigger for a range of chronic conditions, including asthma, hypertension, and low self-esteem5, and is associated with substantial and adverse long-term consequences for overall health, particularly cardiometabolic health6. More than half of children with obesity are likely to become adults with obesity and to develop obesity-related disorders typically observed in adulthood, including type 2 diabetes7. More importantly, obesity-related insulin resistance and abnormalities in glucose metabolism during childhood and adolescence may substantially increase the risk of developing type 2 diabetes and other metabolic diseases during adolescence and later adulthood8–10. The World Health Organization has emphasized that pharmacological interventions face several challenges, including high costs, complications, and counterfeit products11. Therefore, the establishment of healthy lifestyle behaviors, such as regular exercise, remains a primary strategy for the prevention and management of childhood obesity12–14. As a comprehensive exercise modality, combined aerobic and resistance exercise has recently been considered a potentially advantageous approach for managing children and adolescents with overweight or obesity. This intervention may exert synergistic effects on body composition, physical function, and glucose regulation, thereby providing a more comprehensive strategy for improving obesity-related insulin resistance and abnormalities in glucose metabolism15,16. However, although previous meta-analyses have evaluated exercise interventions for insulin resistance-related outcomes in children and adolescents with overweight or obesity, most have either pooled different exercise modalities or compared multiple exercise types at a broader level. Evidence specifically focusing on combined aerobic and resistance exercise as a distinct intervention modality remains less clearly summarized. Therefore, a focused and updated synthesis of this exercise modality is warranted. Therefore, the present study conducted a systematic review and meta-analysis to quantitatively synthesize the effects of combined aerobic and resistance exercise on glucose metabolism and insulin resistance in children and adolescents with overweight or obesity. In addition, subgroup analyses by continent, sex, session duration, and weight status, as well as meta-regression analyses based on age, intervention duration, and exercise frequency, were performed to explore potential sources of heterogeneity. The findings of this study may provide evidence to inform the optimization of exercise interventions and improve metabolic health in children and adolescents with overweight or obesity. Materials and methods This study was conducted in accordance with the Cochrane Handbook for Systematic Reviews of Interventions17 and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses PRISMA 2020 guidelines18. The study protocol was registered in the PROSPERO database (registration number: CRD420261372276). Search strategy A systematic literature search was performed in four electronic databases, including PubMed, Cochrane Library, Web of Science, and Embase, from database inception to March 1, 2026. The search strategy combined database-specific controlled vocabulary terms, where available, with free-text keywords related to children and adolescents, overweight or obesity, combined aerobic and resistance exercise, glucose metabolism, and insulin resistance. Boolean operators were used to combine search terms, and the search strategy was adapted according to the indexing system and search interface of each database. The search was last updated on March 1, 2026. The review focused on published studies indexed in the selected databases, and grey literature and unpublished studies were not systematically searched. No language restriction was applied during the database search. The complete search strategies for all databases are provided in Supplementary Material: Appendix 1. Inclusion and exclusion criteria Studies were considered eligible if they met the following criteria: (1) randomized controlled trials; (2) included children and adolescents with overweight or obesity aged 18 years or younger; (3) defined overweight or obesity in children and adolescents using age- and sex-specific pediatric BMI criteria reported in the original studies, including BMI-for-age percentiles, BMI-for-age z-scores, WHO criteria, CDC criteria, IOTF criteria, or national pediatric BMI reference standards. In percentile-based criteria, overweight is commonly defined as BMI ≥ 85th percentile and obesity as BMI ≥ 95th percentile19–21; (4) included at least one intervention group receiving combined aerobic and resistance exercise; (5) used a control group receiving usual care, usual treatment, or no structured exercise intervention; and (6) reported at least one of the following outcomes: fasting plasma glucose, fasting insulin, or homeostasis model assessment of insulin resistance (HOMA-IR), with sufficient data available for meta-analysis. The exclusion criteria were as follows: (1) non-randomized controlled trials; (2) studies involving special populations, including participants with genetic disorders affecting glucose metabolism, severe metabolic syndrome, or other major comorbidities; (3) studies with intervention types outside the predefined scope, including those involving poorly defined exercise modalities such as sports games or sports competitions; (4) studies including dietary interventions, in order to isolate the independent effects of combined aerobic and resistance exercise on glucose metabolism and insulin resistance; (5) studies that did not report relevant outcomes or provided incomplete data; and (6) duplicate publications or studies with insufficient information. Study selection and data extraction All records identified through the literature search were independently screened by two reviewers (ZH-Q and XY-X). Titles and abstracts were first reviewed to exclude clearly irrelevant studies, after which the full texts of potentially eligible articles were assessed to determine final inclusion. Any disagreement arising during the screening process was resolved through discussion, and if consensus could not be reached, a third reviewer (QZ-L) was consulted for adjudication. For data extraction, two reviewers (ZH-Q and XY-X) independently extracted data in duplicate using a standardized data extraction form. The following information was collected: first author, year of publication, country, participant characteristics (sample size, age, and sex), intervention characteristics (frequency, session duration, and intervention duration of combined aerobic and resistance exercise), control conditions, and outcome measures. The primary outcomes were fasting plasma glucose, fasting insulin, and HOMA-IR. Means and standard deviations for each outcome at baseline and at the end of the intervention were extracted for the subsequent meta-analysis. Risk of bias assessment and certainty of evidence Two reviewers (ZH-Q and XY-X) independently assessed the methodological quality of the included studies using the Cochrane Risk of Bias 2 (RoB 2) tool and classified each study as having “low risk,” “some concerns,” or “high risk” of bias22. The certainty of evidence was evaluated using the GRADE approach23,24, considering risk of bias, inconsistency, imprecision, indirectness, and publication bias. Any disagreements were resolved through discussion. Statistical analysis All statistical analyses were performed using R software (version 4.5.1), with the metafor package used for data processing25. During data extraction, the units of all outcome measures were checked against the original reports. Before quantitative synthesis, fasting plasma glucose was standardized to mmol/L, fasting insulin was standardized to µIU/mL, and HOMA-IR was treated as a unitless index. Outcomes reported in other units were converted to common units before analysis26,27. Mean differences (MDs) with 95% confidence intervals (CIs) were calculated only after unit harmonization within each outcome to estimate the effects of combined aerobic and resistance exercise on glucose metabolism- and insulin resistance-related outcomes in children and adolescents with overweight or obesity. Means and standard deviations for fasting plasma glucose, fasting insulin, and HOMA-IR at baseline and post-intervention were extracted, and the meta-analysis was conducted based on change scores28. Given the potential clinical and methodological differences across studies in terms of participant characteristics and exercise prescriptions, including frequency, intensity, and duration, a random-effects model was used for all main meta-analyses and subgroup analyses. Between-study heterogeneity was assessed using the I² statistic, with I² > 50% generally considered to indicate substantial heterogeneity29. To further explore potential sources of heterogeneity, subgroup analyses were conducted according to continent, sex, weight status, and session duration, and meta-regression analyses were performed using age, intervention duration, and exercise frequency as covariates. Publication bias was assessed by visual inspection of funnel plots in combination with Begg’s test and Egger’s regression test. Sensitivity analyses were conducted using a leave-one-out approach to evaluate the robustness of the pooled results. Results Study selection A total of 1,202 records were identified through database searching. After importation into EndNote 2025, duplicate records were removed. The remaining studies were then screened independently by the reviewers based on the predefined inclusion and exclusion criteria through title and abstract review. Full texts of potentially eligible studies were subsequently retrieved and assessed in detail. Ultimately, 18 studies met the inclusion criteria and were all included in the meta-analysis (Fig. 1). PRISMA flow diagram of study selection process. Characteristics of the included studies A total of 18 randomized controlled trials were included in this study16,30–46, published between 2007 and 2023, involving 682 children and adolescents with overweight or obesity, including 343 participants in the intervention group and 339 in the control group. The included studies were conducted across a relatively broad range of regions, mainly in Korea, the United States, and Iran, with additional studies from Brazil, Switzerland, Spain, and Singapore, indicating a certain degree of geographic representativeness. Participants were generally in childhood or adolescence, with reported mean baseline ages ranging from approximately 8.8 to 15.8 years. Regarding sex distribution, eight studies included mixed-sex populations, five included females only, four included males only, and one study did not clearly report sex distribution. Some studies reported baseline BMI values, which ranged approximately from 28.6 to 30.60 kg/m². All included studies used combined aerobic and resistance exercise as the intervention; however, there was some variation in training dose across studies. Intervention duration ranged from 8 to 20 weeks, with 12 weeks being the most common. Exercise frequency ranged from 2 to 5 sessions per week, with 3 sessions per week being the most frequently used. Session duration ranged from 30 to 90 min, with most studies clustering around 60 min. Overall, the included studies showed a certain degree of heterogeneity in intervention duration, exercise frequency, and session duration. Detailed participant characteristics and intervention protocols are presented in the Supplementary Material: Appendix 2. Risk of bias and certainty of evidence The results of the risk-of-bias assessment are shown in Fig. 2, and the detailed assessments are provided in the Supplementary Material: Appendix 3. Based on the RoB 2 tool, the overall methodological quality of the included studies was predominantly rated as “some concerns.” Among the 18 included studies, 5 were judged to be at low risk of bias, 12 were rated as having some concerns, and 1 was rated as high risk. Across individual bias domains, 8 studies were rated as low risk, 9 as having some concerns, and 1 as high risk for D1 (bias arising from the randomization process); 6 studies were rated as low risk and 12 as having some concerns for D2 (bias due to deviations from intended interventions); 6 studies were rated as low risk and 12 as having some concerns for D3 (bias due to missing outcome data); 6 studies were rated as low risk and 12 as having some concerns for D4 (bias in measurement of the outcome); and 7 studies were rated as low risk and 11 as having some concerns for D5 (bias in selection of the reported result). Overall, the main methodological limitations were related to the randomization process, deviations from intended interventions, and missing outcome data, whereas the risk of bias in outcome measurement was generally low. According to the GRADE assessment, the certainty of evidence was rated as low for fasting plasma glucose and fasting insulin, mainly due to risk of bias and between-study heterogeneity. The certainty of evidence for HOMA-IR was rated as very low, primarily because of risk of bias and extremely high heterogeneity. Overall, although combined aerobic and resistance exercise showed beneficial effects on glucose metabolism and insulin resistance-related outcomes in children and adolescents with overweight or obesity, the certainty of the available evidence remains limited, and these findings should therefore be interpreted with caution. Meta-analysis result Fasting plasma glucose A meta-analysis was performed for 15 studies involving 569 participants (Fig. 3). The results showed that combined aerobic and resistance exercise significantly reduced fasting plasma glucose levels in children and adolescents with overweight or obesity (MD = − 0.19, 95% CI: −0.31 to − 0.07; P = 0.0016). Substantial heterogeneity was observed among the included studies (I² = 76.4%); therefore, further subgroup analyses (Supplementary Material: Appendix 5) and meta-regression analyses (Supplementary Material: Appendix 6) were conducted. Subgroup analyses showed that sex and session duration were significantly associated with improvements in fasting plasma glucose. Specifically, a significant improvement was observed in the female subgroup (MD = − 0.42, 95% CI: −0.73 to − 0.10), and studies with a session duration of ≥ 60 min showed a more pronounced glucose-lowering effect (MD = − 0.26, 95% CI: −0.41 to − 0.11). In contrast, no significant between-subgroup differences were observed according to continent or weight status. Meta-regression analyses showed that age, exercise frequency, and intervention duration were not significantly associated with the effect size for fasting plasma glucose. Risk of bias summary of included studies. Fasting insulin For fasting insulin, 14 studies involving 528 participants were included in the meta-analysis (Fig. 4). The results showed that combined aerobic and resistance exercise significantly reduced fasting insulin levels in children and adolescents with overweight or obesity (MD = − 5.58, 95% CI: −8.18 to − 2.98; P < 0.0001), although substantial heterogeneity was present across studies (I² = 80.4%). Further subgroup analyses (Supplementary Material: Appendix 7) and meta-regression analyses (Supplementary Material: Appendix 8) suggested that continent may influence the intervention effect. Significant improvements were observed in both the Asian subgroup (MD = − 6.84, 95% CI: −9.92 to − 3.76), and the between-subgroup difference across continents was statistically significant (P = 0.0020). Meta-regression further showed that age was significantly associated with the effect size for fasting insulin (β = −1.09, P = 0.03, R² = 35.19%), suggesting that the beneficial effect of exercise on fasting insulin may become more pronounced with increasing age. By contrast, sex, weight status, session duration, exercise frequency, and intervention duration were not significantly associated with the effect size. Forest plot of the effects of combined aerobic and resistance exercise on fasting plasma glucose using a random-effects model. HOMA-IR A meta-analysis of HOMA-IR included 15 comparisons involving 568 participants (Fig. 5). The results showed that combined aerobic and resistance exercise was statistically associated with a reduction in HOMA-IR levels in children and adolescents with overweight or obesity (MD = − 1.83, 95% CI: −3.00 to − 0.66; P = 0.0022), but with very high heterogeneity (I² = 96.9%). Further subgroup analyses (Supplementary Material: Appendix 9) and meta-regression analyses (Supplementary Material: Appendix 10) indicated that continent was significantly associated with improvements in HOMA-IR, with more pronounced effects observed in the North American and Asian subgroups. In contrast, no significant between-subgroup differences were found according to sex, session duration, or weight status. Meta-regression analyses showed that age was significantly associated with the effect size for HOMA-IR (β = −0.46, P = 0.01, R² = 34.68%), suggesting that the beneficial effect of combined aerobic and resistance exercise on HOMA-IR may increase with age. No significant associations were observed for exercise frequency or intervention duration. However, because the certainty of evidence for HOMA-IR was rated as very low, these findings should be interpreted cautiously. Forest plot of the effects of combined aerobic and resistance exercise on fasting insulin using a random-effects model. Forest plot of the effects of combined aerobic and resistance exercise on HOMA-IR using a random-effects model. Sensitivity analysis Sensitivity analyses for fasting plasma glucose, fasting insulin, and HOMA-IR were conducted using a leave-one-out approach (Supplementary Material: Appendix 11). The results showed that sequential omission of any single study did not materially alter the pooled effect estimates for any of the three outcomes, and the direction of the effects remained generally unchanged. These findings indicate that the meta-analysis results were relatively robust and were not driven by any single study. Publication bias Publication bias was assessed using funnel plots in combination with Begg’s test and Egger’s test. No clear evidence of publication bias was detected for fasting plasma glucose or fasting insulin. For fasting plasma glucose, the P values for Begg’s test and Egger’s test were 0.1133 and 0.1429, respectively; for fasting insulin, the corresponding values were 0.1889 and 0.2198. In both cases, the funnel plots were generally symmetrical (Supplementary Material: Appendices 12–13). For HOMA-IR, the funnel plot showed some degree of asymmetry, and Egger’s test suggested possible publication bias (P = 0.0164), although Begg’s test was not statistically significant (P = 0.3223). After applying the trim-and-fill method and imputing seven hypothetical missing studies (Supplementary Material: Appendix 14), the adjusted pooled effect estimate changed from MD = − 1.83 (95% CI: −3.00 to − 0.66) to MD = − 3.65 (95% CI: −4.89 to − 2.41). The direction of the effect remained unchanged before and after adjustment, and both estimates remained statistically significant, suggesting that the overall result for this outcome was still relatively robust. Discussion The findings of this systematic review and meta-analysis showed that combined aerobic and resistance exercise was associated with improvements in fasting plasma glucose and fasting insulin, and may be associated with reductions in HOMA-IR in children and adolescents with overweight or obesity. However, because the certainty of evidence for HOMA-IR was very low and heterogeneity was very high, this finding should be interpreted cautiously. These findings suggest that the potential value of combined aerobic and resistance exercise may extend beyond weight control or physical fitness enhancement and may also have favorable effects on obesity-related abnormalities in glucose metabolism and insulin resistance. Given that childhood and adolescence represent a critical window for the development and progression of metabolic abnormalities, the present study provides further evidence supporting the use of combined aerobic and resistance exercise to promote metabolic health in this population. In addition, the subgroup and meta-regression analyses suggested that the intervention effects may be influenced by factors such as age, sex, session duration, and continent. However, given the substantial heterogeneity across outcomes, these pooled estimates should be interpreted as average effects across heterogeneous populations and exercise protocols rather than definitive effects applicable to all children and adolescents with overweight or obesity. The present findings are generally consistent with previous evidence showing that exercise interventions can improve obesity-related metabolic abnormalities in children and adolescents. Previous meta-analyses have reported that exercise interventions, particularly aerobic exercise, can improve insulin resistance-related markers in children and adolescents with obesity, and that these effects may be more pronounced in adolescents, in studies with longer intervention durations, or under higher training doses47,48. Subsequent studies further indicated that exercise training is generally beneficial for reducing fasting insulin and HOMA-IR in children and adolescents with overweight or obesity49. More importantly, previous systematic reviews and network meta-analyses have suggested that, compared with single-modality exercise, interventions including a resistance component, particularly combined aerobic and resistance exercise, may have greater potential to improve markers of insulin resistance in children and adolescents with obesity48,50. Compared with these earlier studies, the present study provides a more focused and updated synthesis of combined aerobic and resistance exercise as a distinct intervention modality, rather than pooling it with other exercise types. This distinction is clinically relevant because combined training may simultaneously target cardiorespiratory fitness, skeletal muscle function, and glucose disposal, which are all closely related to obesity-related metabolic dysfunction. In addition, the revised analysis applied stricter eligibility criteria by excluding studies with dietary co-interventions, special clinical populations, and potential duplicate samples, and further explored potential heterogeneity by age, sex, session duration, continent, and weight status. Therefore, the present study should be viewed not as a simple repetition of previous exercise meta-analyses, but as a focused synthesis of the metabolic effects of combined aerobic and resistance exercise in children and adolescents with overweight or obesity. Our findings further support the potential benefits of combined training for insulin resistance-related markers and basal glucose homeostasis, while also highlighting the need for cautious interpretation given the heterogeneity and limited certainty of evidence. The more direct evidence provided by this study may be related to the multifaceted regulatory effects of combined aerobic and resistance exercise on adipose tissue, skeletal muscle, and hepatic glucose metabolism. Obesity-related abnormalities in glucose metabolism in children and adolescents are commonly accompanied by adipose tissue dysfunction, ectopic lipid deposition, and reduced insulin sensitivity in skeletal muscle and the liver51,52. Against this background, the aerobic component may help improve oxidative metabolic capacity, reduce visceral and hepatic fat burden, and ameliorate the metabolic milieu associated with obesity53. The resistance component, in turn, may help preserve or increase lean mass, enhance the capacity of skeletal muscle as the primary site of glucose disposal, and promote adaptive changes related to glucose uptake and improved insulin sensitivity54. On the one hand, combined aerobic and resistance exercise may increase glucose uptake and utilization in peripheral tissues; on the other hand, it may also improve hepatic insulin sensitivity and strengthen the suppression of hepatic glucose output, thereby jointly contributing to the maintenance of basal glucose homeostasis55. Within this mechanistic framework, the reduction in fasting insulin suggests a lower insulin secretory burden required to maintain basal glycemic homeostasis. Because HOMA-IR is jointly derived from fasting plasma glucose and fasting insulin, its concurrent improvement further supports the alleviation of insulin resistance. The significant reduction in fasting plasma glucose also indicates that the metabolic benefits of combined aerobic and resistance exercise may not be limited to enhanced insulin sensitivity, but may further extend to the optimization of basal glucose homeostasis itself. Although the present study overall supports the beneficial effects of combined aerobic and resistance exercise on glucose metabolism and insulin resistance in children and adolescents with overweight or obesity, a certain degree of heterogeneity remained across outcomes, suggesting that the intervention effect may be jointly influenced by developmental stage, training dose, sex composition, and population background. First, age was significantly associated with improvements in fasting insulin and HOMA-IR. One possible explanation is that insulin sensitivity physiologically declines during puberty, and this change is closely related to growth hormone, insulin-like growth factor-1, and elevated free fatty acids resulting from increased lipolysis. Accordingly, older adolescents may be at a stage characterized by more pronounced insulin resistance and therefore may have greater room for improvement56,57. It should also be noted that insulin resistance during adolescence is not entirely pathological. Insulin sensitivity physiologically declines during puberty, particularly around mid-puberty, and this transient change is partly related to pubertal hormonal alterations, including increased growth hormone secretion. Therefore, the interpretation of HOMA-IR in children and adolescents should be age-, sex-, and pubertal-stage-specific rather than based on adult cutoff values58. In this context, the age-related findings observed in the present study may partly reflect differences in pubertal maturation rather than chronological age alone. Second, the improvement in fasting plasma glucose was associated with session duration, suggesting that glucose homeostasis may be relatively sensitive to training dose. Previous randomized controlled trials have shown that, at the same exercise intensity, 40 min/day of aerobic training produced greater improvements in insulin resistance and fat mass than 20 min/day. More recent network meta-analytic evidence has further indicated that, in children and adolescents with overweight or obesity, an exercise dose of approximately 900–1200 MET-min per week, which generally corresponds to 2–3 sessions per week of about 60 min each, may be more likely to yield clinically meaningful improvements in insulin resistance48,59. This, to some extent, supports the findings of the subgroup with a session duration of ≥ 60 min in the present study. Third, the more pronounced improvement in fasting plasma glucose observed in the female subgroup should be interpreted cautiously, although it is not without physiological plausibility. Previous studies have shown that sex differences in insulin sensitivity emerge around puberty, and these differences are closely related to fat distribution, lean mass proportion, and pubertal progression. In pediatric obesity, the associations of HOMA-IR with fat distribution and cardiometabolic risk factors also appear to show a degree of sex specificity60,61. Therefore, the sex difference observed in this study may not simply indicate whether boys or girls benefit more, but may instead reflect underlying differences in developmental stage and body composition between the sexes. Given the study-level nature of the subgroup analysis, this finding should be regarded as exploratory rather than evidence of a causal sex-specific response. Finally, the differences observed in fasting insulin and HOMA-IR across continents also suggest that continent-level differences may reflect deeper population heterogeneity. Previous studies have reported that adolescents from different regions may differ in body fat distribution and insulin sensitivity. For example, South Asian adolescents may exhibit higher body fat percentages and poorer insulin sensitivity than their peers at similar BMI levels, indicating that regional differences may not merely reflect geography, but may instead capture differences in ethnic background, lifestyle, and baseline metabolic status62,63. Therefore, continent-level differences should be considered hypothesis-generating rather than direct biological effects. By contrast, exercise frequency and intervention duration were not significantly associated with the outcomes in the present study. However, this finding should be interpreted cautiously, as the included studies were not fully consistent in reporting exercise intensity, total training load, and adherence, and univariable meta-regression is limited in its ability to disentangle the interactions among different prescription components. Limitations Several limitations of this study should be acknowledged. First, although all included studies were randomized controlled trials, varying degrees of heterogeneity were observed across outcomes, particularly for HOMA-IR, which showed very high heterogeneity. This suggests that substantial differences remained in participant characteristics, baseline metabolic status, and exercise prescriptions across studies. Second, although all included interventions were based on combined aerobic and resistance exercise, they were not fully consistent in terms of intervention duration, exercise frequency, session duration, and specific training intensity. Although exercise intensity was additionally extracted and reported in the study characteristics table, it could not be quantitatively examined in subgroup or meta-regression analyses because intensity was inconsistently reported across studies, using different indicators such as HRmax, HRR, VO₂peak, VO₂max, 1RM, or qualitative descriptions. This limited our ability to assess potential intensity-response relationships. The order of aerobic and resistance exercise was also inconsistently reported, limiting interpretation of its potential influence on metabolic outcomes. Such inconsistencies at the prescription level may have affected the comparability of results across studies. Third, the number of studies included in some subgroup and meta-regression analyses was relatively limited. After stratification by continent, sex, and session duration, some subgroups contained only a small number of studies, and the corresponding findings should therefore be interpreted with caution. In addition, the subgroup and meta-regression analyses were exploratory and based on study-level characteristics; therefore, these findings should be interpreted as hypothesis-generating rather than causal evidence. In addition, the methodological quality of some included studies was still predominantly rated as “some concerns,” with the main issues related to the randomization process, deviations from intended interventions, and missing outcome data, which may have influenced the stability of the pooled estimates to some extent. It should also be noted that glucose metabolism and insulin sensitivity in children and adolescents are influenced by multiple factors, including pubertal stage, dietary intake, habitual physical activity, and adherence. Tanner stage was not consistently reported in the included studies, which limited our ability to distinguish physiological pubertal insulin resistance from pathological insulin resistance. Because HOMA-IR should ideally be interpreted using age-, sex-, and pubertal-stage-specific reference values in pediatric populations, the absence of pubertal-stage information reduced the clinical interpretability of the HOMA-IR findings. In addition, although the included studies mainly involved school-age children and adolescents, some participants were at the younger end of the pediatric age range. Because fasting insulin levels and insulin sensitivity vary substantially during childhood, HOMA-IR findings in younger children should be interpreted with particular caution. However, the reporting of these potential confounders was insufficient in the currently available studies, limiting further interpretation of the sources of the intervention effects. Finally, although no clear publication bias was detected for fasting plasma glucose or fasting insulin, Egger’s test suggested possible publication bias for HOMA-IR. Moreover, the overall follow-up duration of the included studies was relatively short. Therefore, the findings of the present study mainly reflect short- to medium-term metabolic benefits, and the long-term sustainability of these effects still requires confirmation from higher-quality trials with longer follow-up periods. Conclusions In summary, combined aerobic and resistance exercise was associated with improvements in fasting plasma glucose and fasting insulin, and may be associated with reductions in HOMA-IR in children and adolescents with overweight or obesity. However, the HOMA-IR finding should be interpreted cautiously because the certainty of evidence was very low and heterogeneity was very high. Subgroup and meta-regression analyses suggested that intervention effects may vary by sex, session duration, age, and continent. These findings support the potential role of combined aerobic and resistance exercise in improving glucose metabolic health in pediatric populations with overweight or obesity, but further high-quality randomized controlled trials with standardized exercise prescriptions and longer follow-up are needed. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Abbreviations Body mass index Confidence interval Grading of recommendations assessment, development and evaluation Homeostasis model assessment of insulin resistance Mean difference Preferred reporting items for systematic reviews and meta-analyses International prospective register of systematic reviews Randomized controlled trial Risk of bias 2 Author contributions Zhihao Qi and Qinzheng Liu designed the study. Zhihao Qi, Xianyang Xin, and Xiangyun Xu performed the literature screening and data extraction. Zhihao Qi, Xianyang Xin, and Qinzheng Liu assessed study quality and conducted the statistical analyses. Yongqing Guo contributed to data interpretation and manuscript revision. Zhihao Qi drafted the manuscript. Qinzheng Liu critically revised the manuscript and approved the final version. All authors read and approved the final manuscript. Data availability The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Code availability No custom software, new algorithm, or bespoke computational tool was developed in this study. All statistical analyses were conducted using standard functions and packages in R software. The analysis code can be made available from the corresponding author upon reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. No custom software, new algorithm, or bespoke computational tool was developed in this study. All statistical analyses were conducted using standard functions and packages in R software. The analysis code can be made available from the corresponding author upon reasonable request.