Exploring Metabolic Signatures: Unraveling the Association with Obesity in Children and Adolescents

Academic Editor: Andrea Vania Received: 15 April 2025 Revised: 21 May 2025 Accepted: 23 May 2025 Published: 28 May 2025 Citation: Koutaki, D.; Stefanou, G.; Genitsaridi, S.-M.; Ramouzi, E.; Kyrkili, A.; Kontogianni, M.D.; Kokkou, E.; Giannopoulou, E.; Kassari, P.; Charmandari, E. Exploring Metabolic Signatures: Unraveling the Association with Obesity in Children and Adolescents. Nutrients 2025,17, 1833. https://doi.org/10.3390/ nu17111833 Copyright: © 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/ licenses/by/4.0/). Systematic Review Exploring Metabolic Signatures: Unraveling the Association with Obesity in Children and Adolescents Diamanto Koutaki 1, Garyfallia Stefanou 2, Sofia-Maria Genitsaridi 1, Eleni Ramouzi 1, Athanasia Kyrkili 3, Meropi D. Kontogianni 3, Eleni Kokkou 1, Eleni Giannopoulou 1, Penio Kassari 1,4 and Evangelia Charmandari 1,4,* 1Center for the Prevention and Management of Overweight and Obesity in Childhood and Adolescence, Division of Endocrinology, Metabolism and Diabetes, First Department of Pediatrics, National and Kapodistrian University of Athens Medical School, ‘Aghia Sophia’, Children’s Hospital, 11527 Athens, Greece; [email protected] (D.K.); [email protected] (S.-M.G.); [email protected] (E.R.); [email protected] (E.K.); [email protected] (E.G.); [email protected] (P.K.) 2ECONCARE—Health Research & Consulting, 11528 Athens, Greece; [email protected] 3Department of Nutrition and Dietetics, School of Health Sciences and Education, Harokopio University of Athens, 17671 Athens, Greece; [email protected] (A.K.); [email protected] (M.D.K.) 4Division of Endocrinology and Metabolism, Center of Clinical, Experimental Surgery and Translational Research, Biomedical Research Foundation of the Academy of Athens, 11527 Athens, Greece *Correspondence: [email protected]; Tel.: +30-213-2013-384 Abstract: Background: Childhood obesity is a growing global health concern. Metabolomics, the comprehensive study of metabolites within biological systems, offers a powerful approach to better define the phenotype and understand the complex biochemical alterations associated with obesity. The aim of this systematic review was to summarize current knowledge in the field of metabolomics in childhood obesity and to identify metabolic signatures or biomarkers associated with overweight/obesity (Ov/Ob) and Metabolically Unhealthy Obesity (MUO) in children and adolescents. Methods: We performed a systematic search of Medline and Scopus databases according to PRISMA guidelines. We included only longitudinal prospective studies or randomized controlled trials with ≥ 12 months of follow-up, as well as meta-analyses of the above that assessed the relation between metabolic signatures related to obesity and Body Mass Index (BMI) or other measures of adiposity in children and adolescents aged 2–19 years with overweight or obesity. Initially, 595 records were identified from PubMed and 1565 from Scopus. After removing duplicates and screening for relevance, 157 reports were assessed for eligibility. From the additional search, 75 new records were retrieved, of which none were eligible for our study. Finally, 7 reports were included in the present systematic review (4 reporting on Ov/Ob and 4 on MUO). Results: The presented studies suggest that the metabolism of amino acids and lipids is primarily affected by childhood obesity. Metabolites like glycoprotein acetyls, the Apolipoprotein B/Apolipoprotein A-1 ratio, and lactate have emerged as potential biomarkers for insulin resistance and metabolic syndrome, highlighting their potential value in clinical applications. Conclusions: There is a need for future longitudinal studies to assess metabolic changes over time, interventional studies to evaluate the efficacy of therapeutic strategies, and large-scale population studies to explore metabolic diversity across different demographics. Our findings reveal specific biomarkers in the amino acid and lipid pathway that may serve as early indicators of childhood obesity and its associated cardiometabolic complications. Keywords: metabolomics; metabolic signatures; metabolic biomarkers; childhood obesity Nutrients 2025,17, 1833 https://doi.org/10.3390/nu17111833 Nutrients 2025,17, 1833 2 of 18 1. Introduction Obesity has emerged as a significant global health issue, with its prevalence having nearly tripled from 1975 to 2016 [1]. According to the World Health Organization (WHO), approximately 60% of adults in Europe will be overweight or obesity in 2022 [ 2 ]. This alarming trend not only poses immediate health risks but also predisposes affected subjects to long-term health complications, such as hypertension, left ventricular hypertrophy, insulin resistance and diabetes mellitus type 2 (DM2), metabolic dysfunction-associated steatotic liver disease (MASLD), as well as mental health issues and cancer [ 3 ]. Furthermore, during the last decade, obesity with or without metabolic aberrations, commonly termed Metabolically Unhealthy Obesity (MUO) or Metabolically Healthy Obesity (MHO), respectively, has been extensively investigated [ 4 ]. Metabolically Unhealthy Obesity (MUO) refers to subjects with obesity who exhibit metabolic abnormalities, such as insulin resistance, elevated blood pressure, dyslipidemia (elevated triglycerides and low HDL cholesterol concentrations), and chronic inflammation. Unlike metabolically healthy obesity (MHO), where subjects have excess body fat but normal metabolic profiles, MUO is strongly associated with an increased risk of DM2, cardiovascular disease, and non-alcoholic fatty liver disease (NAFLD). The intricate interplay of genetic, epigenetic, environmental, and lifestyle factors contributes to the multifaceted nature of this epidemic [ 3 ]. In the quest for a better understanding of the underlying mechanisms and potential interventions, metabolomics has emerged as a powerful tool, offering insights into the metabolic alterations associated with childhood obesity [5,6]. Metabolomics is a field of study within the broader discipline of systems biology, which focuses on the comprehensive analysis of small molecules or metabolites (<1500 KDa) in a biological sample [ 7 , 8 ]. Metabolites are the end products of cellular processes, and their concentrations can provide insights into the biochemical pathways and physiological status of an organism at a specific point in time. It is particularly useful in understanding the dynamic responses of biological systems to various alterations, including genetic, epigenetic, or protein-level modifications, exposure to environmental factors (physical exercise, diet, and microbiome) and diseases, and helps bridge the gap between genotype and phenotype. The primary goal of metabolomics is to profile and quantify the complete set of metabolites present in a biological sample, such as blood, urine, or tissues. This profiling involves the use of advanced analytical techniques, such as mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectroscopy, coupled with various chromatographic separations to identify and quantify the diverse array of metabolites [ 9 ]. This way, the systematic study of small molecules within biological systems may help us gain insight into the metabolic changes associated with obesity, such as adipocyte-related inflammation and insulin resistance [ 10 ]. The in-depth study of the unique metabolic fingerprints associated with obesity may help us identify potential biomarkers and altered metabolic pathways, and discover novel therapeutic targets. Numerous studies have underscored the utility of metabolomics in elucidating the complex interplay between genetic predisposition, dietetic habits, gut microbiome, and environmental factors in the pathogenesis of obesity in adults [ 11 ]. However, to the best of our knowledge, few studies have been conducted in children and adolescents [ 12 ]. Metabolic signatures may differ in early life, given that children do not usually receive medical treatment for obesity. Therefore, metabolomic profiling in children and adolescents will not only facilitate the identification of potential biomarkers for the prevention and management of childhood obesity and its associated complications, but it will also unravel novel therapeutic targets. Nutrients 2025,17, 1833 3 of 18 The aim of this systematic review was to summarize the current knowledge on metabolomics, childhood obesity, and MUO, and to identify metabolic signatures or biomarkers associated with obesity in children and adolescents, thereby offering a comprehensive analysis of studies that employ metabolomic approaches. Through critical examination of the literature, this review aims to gain a better understanding of the metabolic intricacies associated with childhood obesity and inform the direction of future research and therapeutic strategies. 2. Materials and Methods 2.1. Study Design This systematic literature review (SLR) was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol [ 13 ]. The objectives were formulated using the PICO/PECO (Population, Interventions/Exposure, Comparators, Outcomes) framework (Table 1). The review was registered in the International Prospective Register of Ongoing Systematic Reviews (PROSPERO 2023 CRD42023494461; https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD4 2023494461, accessed on 29 December 2023). Table 1. PICO/PECO framework for study selection on metabolomic biomarkers and childhood obesity risk. Variable Definition Population Children and adolescents aged 2–19 years Exposure/Intervention Metabolomics, metabolic signatures, and metabolic biomarkers Comparator No intervention, any intervention, or standard care The absence of the exposure or a different level of exposure Outcome Ov/Ob and MUO risk Association of metabolic signatures/biomarkers with obesity/adiposity/metabolic disorders/endocrine disorders 2.2. Eligibility Criteria The review included longitudinal prospective studies and randomized controlled trials (RCTs), with a minimum of 12 months of follow-up, and meta-analyses of the above in order to ensure a better quality of methodological design, which would also allow etiological assumptions. The studies examined the metabolic signatures related to obesity, Body Mass Index (BMI), or/and other measures of adiposity and MUO in children and adolescents aged 2–19 years with overweight or obesity. The language was restricted to English and the geographic location included only Western countries (Europe, USA, Canada, and Oceania) that share similar socioeconomic, physical, and dietary environments. The inclusion and exclusion criteria are shown in Table 2. Table 2. Inclusion and exclusion criteria. Parameter Inclusion Criteria for All Domains Exclusion Criteria for All Domains Participants Human subjects Animals Human subjects with monogenic disorders (e.g., MC4R deficiency, leptin deficiency, etc.), syndromic forms of obesity (e.g., Prader–Willi, Alstrom syndrome, etc.), or subjects receiving medication known to affect weight (antidepressants, antiepileptics, antipsychotics, mood stabilizers, antimanic agents, and corticosteroids) Nutrients 2025,17, 1833 4 of 18 Table 2. Cont. Parameter Inclusion Criteria for All Domains Exclusion Criteria for All Domains Age 2 to 19 years old <2 years old and >19 years old Article type Peer-reviewed journal articles Letters, editorials, study or review protocols, pre-prints Study area Europe, USA, Canada, Oceania Asia, Africa, South America Study design Longitudinal prospective studies, randomized controlled trials with ≥12 months of follow-up, and meta-analyses of the above Cross-sectional studies, controlled experiments, in vitro studies, in vivo animal studies, in silico studies, and scoping reviews Time of publication 1 January 2013–3 July 2024 for original publications and 1 January 2018–3 July 2024 for meta-analyses Original publications prior to 31 December 2012 and meta-analyses prior to 31 December 2017 Language English Non-English 2.3. Literature Search A comprehensive literature search was conducted using PubMed and Scopus databases for studies published from 1 May 2023 to 16 September 2023. An additional data search was performed on 3 July 2024 to update the results, retrieving studies published after 16 September 2023. The search strategy included a complex string of keywords related to metabolic biomarkers, obesity, adiposity, and associated metabolic and endocrine disorders in children and adolescents. The detailed search strings used for MEDLINE (PubMed) and Scopus are presented in the Supplementary Materials, File S1. 2.4. Study Selection Two independent researchers (GS and DK) screened the records identified from the databases. In instances of disagreement, a third researcher (EC) conducted a final review. 2.5. Data Extraction, Outcomes, and Data Synthesis Relevant data from eligible studies were extracted, including publication details, study design, sample size, participant characteristics, metabolic signatures, biomarkers assessed, and outcomes related to obesity and metabolic disorders. The primary outcomes assessed were obesity and MUO in pediatric populations. 2.6. Validity Assessment All included studies were assessed for risk of bias using the Risk Of Bias In Nonrandomized Studies—of Exposures (ROBINS-E) tool [ 14 ]. The risk of bias for each study was evaluated across seven domains: confounding (D1), measurement of exposure (D2), selection of participants (D3), post-exposure interventions (D4), missing data (D5), measurement of outcomes (D6), and selection of reported results (D7). Three of the included studies were post hoc analyses of participants who underwent a lifestyle intervention (“Obeldicks”) within a non-randomized controlled trial [ 15 , 16 ] or a double-blind, randomized intervention trial [ 17 ]. Since the exposure of interest (metabolites) was not actively assigned, ROBINS-E was deemed the most appropriate tool for assessing the risk of bias in these studies. Nutrients 2025,17, 1833 5 of 18 2.7. Data Management and Synthesis Data were managed using Mendeley and Excel. Data extraction forms were piloted and refined to ensure consistency and accuracy. Discrepancies between reviewers were resolved through discussion and consensus. The extracted data were synthesized to provide a comprehensive analysis of the metabolic signatures associated with childhood obesity and MUO. The synthesis involved qualitative analyses to summarize the findings and identify potential biomarkers and therapeutic targets. The characteristics of the included studies, e.g., study design, country, sample size, age, follow-up period, methodology, key metabolites identified, and reported outcome are presented in Table 3. Table 3. Characteristics of the included studies. Study Country Study Design Sample Size Age, Mean ±SD Follow-Up Period Methodology Key Metabolites Identified Singh et al., 2023 [18] USA Longitudinal cohort study (Buckeye Teen Health Study) 81 (100% males) 16.08 ±1.20 years 1 year UPLC-QTOFMS (urine) Glycylproline, 3’-Sialyllactose, Formiminoglutamic acid, 4-hydroxyproline, Citrulline, Inosine Mansell et al., 2022 [19] Australia Longitudinal cohort study (COBRA cohort) 98 (52% males) 10.3 ±3.5 years 5 years NMR (serum) XL-VLDL-L, L-VLDL-L, S-VLDL-L, ApoB/ApoA1, VLDL-C, MUFAs, MUFAs%, alanine, phenylalanine, tyrosine, pyruvate, glycoprotein acetyls, HDL-C, LA%, Omega-6%, PUFAs, Acetoacetate, 3-hydroxybutyrate Reinehr et al., 2014 [15] Germany Post hoc analysis of participants who underwent a lifestyle intervention (“Obeldicks”) within a non-randomized controlled trial 160 (61.3% males) 11 ±2 years 1 year HPLC-MS (serum) Glutamine, methionine, LPCaC18:1, LPCaC18:2, LPCaC20:4, PCaeC36:2 Hellmuth et al., 2019 [17] Europe (multi) Post hoc longitudinal analysis of biomarker changes over 2.5 years in participants from the CHOP study, a double-blind, randomized intervention trial 396 (50% males) 5.5 ±0.07 years 2.5 years UPLC-QTOFMS (serum) Free carnitine, SM 32:2, SM 34:2, Carn 3:0 Ojanen et al., 2021 [20] Finland Longitudinal cohort study 396 (0% males) 11.2 ±0.4 years 7.5 years NMR (serum) ApoB/ApoA ratio, GlycAs Nutrients 2025,17, 1833 6 of 18 Table 3. Cont. Study Country Study Design Sample Size Age, Mean ±SD Follow-Up Period Methodology Key Metabolites Identified Hellmuth et al., 2016 [16] Germany Post hoc analysis of participants who underwent a lifestyle intervention (“Obeldicks”) within a non-randomized controlled trial 80 (45% males) 11.5 ±2.4 years 1 year HPLC-MS (serum) Acylcarnitines, amino acids Hosking et al., 2019 [21] UK, Switzerland Longitudinal cohort study (EarlyBird cohort) 190 [Study 1: 40 (50% males); Study 2: 150 (70% males)] 4.8–5.1 years Study 1: 9 years; Study 2: 11 years 1H NMR (serum) Amino acids, lipids, lactate 2.8. Ethical Considerations Since this study is a systematic review, ethical approval was not required. However, ethical standards were maintained throughout the review process, ensuring the integrity and accuracy of the findings. 3. Results 3.1. Characteristics of Included Studies Initially, 595 records were identified from PubMed and 1565 from Scopus. After removing duplicates and screening for relevance, 175 reports were assessed for eligibility. Ultimately, 7 (4 longitudinal and 3 post hoc analyses of interventional studies) reports were included in the review. From the additional search, 124 new records were retrieved, from which none were eligible for our study. The flow diagram is presented in Figure 1. Seven reports, which were derived from six studies, met the inclusion criteria and were included in this systematic review. These studies were conducted in various countries, including the USA [ 18 ], Germany [ 15 – 17 ], Belgium [ 17 ], Italy [ 17 ], Poland [ 17 ], Spain [ 17 ], Australia [ 19 ], UK [ 21 ], Switzerland [ 21 ] and Finland [ 20 ]. The included studies were primarily longitudinal cohort studies [ 18 – 21 ] and post hoc analyses of intervention studies [ 15 – 17 ], with follow-up periods ranging from 1 year [ 15 , 16 , 18 ] to 11 years [ 21 ]. The intervention part included lifestyle recommendations regarding physical activity, nutrition, and behavioral therapy for the children and their families. The sample size of these studies varied significantly, from 40 to 396 participants, and the age at baseline ranged from 2 to 19 years. Serum and urine samples were used for the assessment of metabolomic biomarkers. The included studies examined the association between metabolic profiles and obesityrelated outcomes using various approaches (Table 4). Singh et al. (2023) explored the metabolic features associated with increased BMI at one-year follow-up [ 18 ]. Mansell et al. (2022) explored the association between changes in BMI and metabolomic profiles over a 5.5-year follow-up period [ 19 ]. Reinehr et al. (2014) assessed the metabolite changes in children with obesity who underwent a lifestyle intervention and compared those with substantial weight loss to those without substantial weight loss [ 15 ]. Hellmuth et al. (2019) used metabolite concentrations at 5.5 years to predict BMI z-scores at age 8 in the CHOP study [ 17 ]. Ojanen et al. (2021) developed a standardized risk score for Nutrients 2025,17, 1833 7 of 18 metabolic syndrome (MetS), having incorporated metabolic and cardiovascular parameters that confer cardiometabolic risk [ 20 ]. Hellmuth et al. (2016) investigated the association between metabolite changes and HOMA-IR over a one-year lifestyle intervention (physical activity, nutrition education, and behavior therapy) [ 16 ]. Hosking et al. (2019) examined the relation between individual metabolites and insulin resistance (HOMA-IR) in healthy children, taking into account the effects of age, BMI, growth, puberty, adiposity, and physical activity [ 21 ]. This study included a pilot phase to identify metabolically distinct profiles related to insulin resistance and a follow-up phase extending the analysis to the age of 16 years to validate the findings. Records initially retrieved: PubMed (n = 595) Scopus (n = 1,565) Additional records retrieved on 2 nd search (n = 124) Records removed before screening: Duplicate records removed (n = 686; of those 49 were from the 2 nd search) Records screened: (n = 1598) Records excluded: (n = 1422) Reports sought for retrieval: (n = 176) Reports not retrieved: (n = 1) Reports assessed for eligibility: (n = 175) Reports excluded: 168 Other study designs (n = 47) Other objectives (n = 53) Other age groups (n = 19) Asia, Africa, South America (n = 19) Letters, reviews, etc. (n = 17) <12 months of follow-up (n = 12) Animals (n=1) Studies included in review: n = 6 Outcome: Ov/Ob (n = 4) Outcome: MUO (n = 4) Reports of included studies*: n = 7 Identification of studies Identification Screening Included Eligibility Figure 1. PRISMA flow diagram. Notes: Ov/Ob = overweight/obesity; MUO = Metabolically Unhealthy Obesity. * Two reports were referred to a single study. Nutrients 2025,17, 1833 8 of 18 Table 4. Outcomes of interest, statistical analysis, and results of included studies. Author, Year (Reference) Outcomes Statistical Analysis Results Singh et al., 2023 [18] Significant metabolic features associated with positive change in BMI at 1-year follow-up. Estimate (95% CI) based on a stratified linear regression model (age, race, BMI z-score, and total energy intake). Glycylproline: −0.018 (−0.029, 0.007) p= 0.002, 3’-Sialyllactose: 0.009 (0.002, 0.016) p= 0.006, formiminoglutamic acid: 0.016 (0.004, 0.028) p= 0.008, glycylproline: −0.014 (−0.025, 0.003) p= 0.01, 4-hydroxyproline: 0.016 (0.003, 0.03) p= 0.016, Citrulline: 0.01 (0.002, 0.018) p= 0.013, 4-Vinylsyringol: −0.01 (−0.02, 0.001) p= 0.022, Citrulline: 0.012 (0.001, 0.023) p= 0.025, Inosine: 0.005 (0.0004, 0.01) p= 0.03, Mansell et al., 2022 [19] Association of change in BMI from baseline to the end of follow-up (5.5 years) with the change in metabolomic profiles. Coefficients (95% CI) [Benjamini–Hochberg adjusted p-value] of the change in log concentrations of metabolites in SD units decrease in BMI over time per unit (kg/m 2) from linear regression models adjusted for age at each time point and sex. lipoprotein subclasses: XL-VLDL-L: −0.038 (−0.066 to −0.01), p= 0.04; L-VLDL-L: −0.038 (−0.066 to −0.01), p= 0.04; S-VLDL-L: −0.039 (−0.071 to −0.008), p= 0.05; Apolipoproteins: ApoB/ApoA1: − 0.046 ( − 0.073 to − 0.019), p= 0.01; cholesterols: VLDL-C: −0.035 (−0.062 to −0.008), p= 0.05; HDL-C: 0.045 (0.011 to 0.08), p= 0.04; HDL2-C: 0.049 (0.016 to 0.082), p= 0.02; fatty acids: unsaturation: 0.059 (0.022 to 0.097), p= 0.02; MUFAs: −0.041 (−0.068 to −0.014), p= 0.02; LA%: 0.065 (0.03 to 0.101), p= 0.01; Omega-6%: 0.069 (0.034 to 0.103), p= 0.003; PUFAs%: 0.065 (0.03 to 0.1), p= 0.01; MUFAs%: −0.061 (−0.094 to −0.028), p= 0.01; amino acids: alanine: −0.072 ( − 0.105 to − 0.04), p= 0.002; phenylalanine: − 0.069 (−0.102 to −0.037), p= 0.002; tyrosine: −0.068 (−0.099 to −0.037), p= 0.002; glycerides and phospholipids: total triglycerides: −0.043 (−0.069 to −0.016), p= 0.02; VLDL-TGs: -0.042 (−0.07 to −0.015), p= 0.02; TG/PG: −0.052 (−0.081 to −0.023), p= 0.01; glycolysis-related metabolites: pyruvate: −0.077 (−0.114 to −0.039), p= 0.002; ketone bodies: Acetoacetate: 0.065 (0.021 to 0.109), p= 0.02; 3-hydroxybutyrate: 0.066 (0.018 to 0.113), p= 0.04; inflammation: glycoprotein acetyls: −0.063 (−0.092 to −0.035), p= 0.002. Additional adjustment for pubertal status confirmed statistically significant associations for fatty acids: LA%, PUFAs%, and MUFAs%; amino acids: alanine, phenylalanine, and tyrosine; glycolysis-related metabolites: pyruvate; and inflammation: glycoprotein acetyls. Reinehr et al., 2014 [15] Change in metabolites between groups (children with obesity with substantial weight loss and children with obesity without weight loss; all underwent a lifestyle intervention). The 14 metabolites [glutamine, methionine, proline, nine phospholipids (PCaeC34:1, C34:2, C34:3, C36:2, C36:3, C38:2, LPCaC18:1, C18:2, and C20:4), and two acylcarnitines (C12:1 and C16:1)] were compared between baseline and 1-year follow-up. The 14 metabolites did not change significantly in children without weight loss. In children with substantial weight loss, glutamine [mean (SD) at baseline: 567 (120), follow-up: 588 (102), p= 0.013], methionine [mean (SD) at baseline: 27 (6), follow-up: 29 (6), p= 0.026], LPCaC18:1 [mean (SD) at baseline: 10 (2.8), follow-up: 10.9 (3), p= 0.003], LPCaC18:2 [mean (SD) at baseline: 12.3 (5.2), follow-up: 13.5 (5.2), p= 0.035], LPCaC20:4 [mean (SD) at baseline: 19.6 (8.2), follow-up: 21.7 (7.7), p= 0.011] and PCaeC36:2 [mean (SD) at baseline: 4.5 (1.7), follow-up: 4.8 (1.4), p= 0.026] increased significantly, while the other eight metabolites did not change significantly. Nutrients 2025,17, 1833 9 of 18 Table 4. Cont. Author, Year (Reference) Outcomes Statistical Analysis Results Hellmuth et al., 2019 [17] Researchers used the metabolite concentrations at 5.5 years to predict the BMI z-score at 8 years of age in the CHOP study. Linear regression models adjusted for child age and gender. Plasma levels of free carnitine (p= 6.17 ×10−6), SM 32:2 (p= 2.16 × 10 −4 ), SM 34:2 (p= 3.09 × 10 −4 ) and Carn 3:0 (p= 4.09 ×10−2) were significantly positively associated with the BMI z-score at 8 years of age. However, after adjusting for the BMI z-score at 5.5 years, no metabolite reached the significance level. Regarding HOMA, glutamine at age 5.5 years was significantly negatively associated (p= 0.013/0.003) with HOMA indices at 8 years in both the unadjusted and adjusted linear models. NEFAs 26:1 (p= 0.012/0.015), 26:2 (p= 0.002/0.01), and 26:3 (p= 0.009/0.015) at age 5.5 years were significantly positively associated with HOMA indices at 8 years in both the unadjusted and adjusted linear models. Only serine levels remained significantly associated with HOMA in the adjusted model (p= 0.032). Ojanen et al., 2021 [20] To assess cardiometabolic risk, a standardized continuously distributed variable for clustered metabolic risk (MetS score) was constructed. The risk score was calculated by standardizing and then summing the following continuously distributed metabolic traits: mean arterial pressure ([(2 × diastolic blood pressure) + systolic blood pressure]/3); abdominal fat mass; fasting plasma glucose; serum HDL cholesterol x −1; and fasting serum triglyceride z-score. The z-scores for each variable and MetS scores were calculated separately for each time point. A higher score indicated a higher cardiometabolic risk. Regression analysis with MetS score as the dependent variable and metabolic biomarkers identified by LASSO as independent variables, after Bonferroni correction for multiple tests. Baseline ApoB/ApoA ratio and GlycAs positively predicted while L-HDL-PLs negatively predicted 7.5-year Mets (r = 0.471 and p< 0.0001; r = 0.400 and p= 0.0005; and r = − 0.465 and p< 0.0001, respectively, p: adjusted for multiple comparisons by Bonferroni). And 2-year ApoB/ApoA ratio and GlycAs positively predicted and L-HDL-PLs negatively predicted 7.5-year Mets (r = 0.449 and p< 0.0001; r = 0.440 and p< 0.0001; and r = − 0.445 and p< 0.0001, respectively, p: adjusted for multiple comparisons by Bonferroni) only. ApoB/ApoA ratio, GlycAs, and L-HDL-PLs remained significant predictors of MetS score (p< 0.0001 for all). These associations were also robust to multi-covariate adjustment, including insulin, leptin, adiponectin, sex steroids, IGF-1, physical activity, and energy yield nutrient intakes. Hellmuth et al., 2016 [16] Association of changes in metabolite concentrations with change in HOMA over the one-year intervention. Change was defined as the relative change over the one-year intervention, with estimates reported alongside 95% confidence intervals (CIs). To assess the association between metabolites and markers of insulin resistance, a two-step robust regression approach was used. First, metabolite levels were adjusted for BMI using ageand sex-adjusted robust regression (M-estimator with Huber bi-square weighting). The residuals from this model were then regressed on the relative change in HOMA over the intervention period, using robust regression to minimize the influence of outliers. All: Carn C0 1.10 [0.29; 1.90] p= 0.008, Carn C6:1-DC −0.33 [−0.59; −0.06] p= 0.015, Carn C6-oxo −0.24 [−0.43; −0.05] p= 0.014, Pro 0.81 [0.19; 1.40] p= 0.011; ratio of Carn C5/Carn C6-oxo 0.24 [0.07; 0.41] p= 0.007, ratio of Carn C6-oxo/xLeu −0.19 [−0.34; −0.03] p= 0.016, Tyr 0.79 [0.17; 1.40] p= 0.015; weight loss: AAA sum 1.04 [0.29; 1.80] p= 0.009, Carn C0 1.71 [0.88; 2.50] p< 0.001, Carn C3 0.49 [0.03; 0.96] p= 0.036; Carn C6:1-DC −0.29 [−0.47; −0.10] p= 0.003; Carn C6-oxo −0.21 [−0.35; −0.08] p= 0.003, Pro 0.72 [0.11; 1.30] p= 0.023, ratio of Carn C4/Carn C5-oxo 0.48 [0.18; 0.77] p= 0.0030, ratio of Carn C5/Carn C6-oxo 0.22 [0.08; 0.35] p= 0.002, ratio of Carn C6:1-DC/Carn C5:1 − 0.22 [ − 0.41; − 0.03] p= 0.024, ratio of Carn C6-oxo/xLeu −0.15 [−0.25; −0.04] p= 0.007, Trp 1.13 [0.14; 2.10] p= 0.027, Tyr 1.09 [0.51; 1.70] p= 0.001, Val 0.73 [0.07; 1.40] p= 0.033; no weight loss: ratio of Carn C5/Carn C6-oxo 0.29 [0.02; 0.57] p= 0.041. Nutrients 2025,17, 1833 16 of 18 stringent methodological approaches in future studies to enhance the reliability of the conclusions drawn. As such, continued exploration of metabolomic profiles in childhood obesity is warranted, particularly in pediatrics, to develop targeted interventions and prevent the long-term consequences of this condition. Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/nu17111833/s1. File S1. Search strings used for MEDLINE (PubMed) and Scopus. Author Contributions: Conceptualization, E.C.; methodology, D.K., G.S., A.K., M.D.K., P.K. and E.C.; software, G.S., validation, P.K. and E.C.; formal analysis, D.K., G.S., P.K. and E.C.; investigation, D.K., G.S. and P.K.; resources, P.K. and E.C.; data curation, D.K., G.S. and P.K.; writing—original draft preparation, D.K., G.S. and P.K.; writing—review and editing, D.K., G.S., S.-M.G., E.R., A.K., M.D.K., E.K., E.G., P.K. and E.C.; visualization, E.C.; supervision, P.K. and E.C.; project administration, P.K. and E.C.; funding acquisition, E.C. All authors have read and agreed to the published version of the manuscript. Funding: The work leading to these results has received funding from the HORIZON European Research and Innovation Action project under Grant Agreement No. 101080718. The project is entitled “Multi-Pillar Framework for children Anti-Obesity Behavior building on an EU biobank, Micro Moments and Mobile Recommendation Systems”, Acronym: BIO-STREAMS, https://www. bio-streams.eu/. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Conflicts of Interest: The authors declare no conflicts of interest. References 1. World Health Organization. Obesity and Overweight. Available online: https://www.who.int/news-room/fact-sheets/detail/ obesity-and-overweight (accessed on 5 February 2024). 2. WHO. European Regional Obesity Report. 2022. Available online: https://www.who.int/europe/publications/i/item/97892890 57738 (accessed on 5 February 2024). 3. Kumar, S.; Kelly, A.S. Review of Childhood Obesity: From Epidemiology, Etiology, and Comorbidities to Clinical Assessment and Treatment. Mayo Clin. Proc. 2017,92, 251–265. [CrossRef] [PubMed] 4. Mathew, H.; Farr, O.M.; Mantzoros, C.S. Metabolic Health and Weight: Understanding metabolically unhealthy normal weight or metabolically healthy obese patients. Metabolism 2016,65, 73–80. [CrossRef] [PubMed] 5. Hivert, M.F.; Perng, W.; Watkins, S.M.; Newgard, C.S.; Kenny, L.C.; Kristal, B.S.; Patti, M.E.; Isganaitis, E.; DeMeo, D.L.; Oken, E.; et al. Metabolomics in the developmental origins of obesity and its cardiometabolic consequences. J. Dev. Orig. Health Dis. 2015, 6, 65–78. [CrossRef] [PubMed] 6. Regan, J.A.; Shah, S.H. Obesity Genomics and Metabolomics: A Nexus of Cardiometabolic Risk. Curr. Cardiol. Rep. 2020,22, 174. [CrossRef] 7. Bujak, R.; Struck-Lewicka, W.; Markuszewski, M.J.; Kaliszan, R. Metabolomics for laboratory diagnostics. J. Pharm. Biomed. Anal. 2015,113, 108–120. [CrossRef] 8. Schrimpe-Rutledge, A.C.; Codreanu, S.G.; Sherrod, S.D.; McLean, J.A. Untargeted Metabolomics Strategies-Challenges and Emerging Directions. J. Am. Soc. Mass. Spectrom. 2016,27, 1897–1905. [CrossRef] 9. Wang, R.; Li, B.; Lam, S.M.; Shui, G. Integration of lipidomics and metabolomics for in-depth understanding of cellular mechanism and disease progression. J. Genet. Genom. 2020,47, 69–83. [CrossRef] 10. Ouchi, N.; Parker, J.L.; Lugus, J.J.; Walsh, K. Adipokines in inflammation and metabolic disease. Nat. Rev. Immunol. 2011, 11, 85–97. [CrossRef] 11. Rangel-Huerta, O.D.; Pastor-Villaescusa, B.; Gil, A. Are we close to defining a metabolomic signature of human obesity? A systematic review of metabolomics studies. Metabolomics Off. J. Metabolomic Soc. 2019,15, 93. [CrossRef] Nutrients 2025,17, 1833 17 of 18 12. Handakas, E.; Lau, C.H.; Alfano, R.; Chatzi, V.L.; Plusquin, M.; Vineis, P.; Robinson, O. A systematic review of metabolomic studies of childhood obesity: State of the evidence for metabolic determinants and consequences. Obes. Rev. Off. J. Int. Assoc. Study Obes. 2022,23 (Suppl. S1), e13384. [CrossRef] 13. Shamseer, L.; Moher, D.; Clarke, M.; Ghersi, D.; Liberati, A.; Petticrew, M.; Shekelle, P.; Stewart, L.A.; The PRISMA-P Group. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: Elaboration and explanation. BMJ 2015,350, g7647. [CrossRef] [PubMed] 14. Higgins, J.P.; Morgan, R.L.; Rooney, A.A.; Taylor, K.W.; Thayer, K.A.; Silva, R.A.; Lemeris, C.; Akl, E.A.; Bateson, T.F.; Berkman, N.D.; et al. A tool to assess risk of bias in non-randomized follow-up studies of exposure effects (ROBINS-E). Environ. Int. 2024, 186, 108602. [CrossRef] [PubMed] 15. Reinehr, T.; Wolters, B.; Knop, C.; Lass, N.; Hellmuth, C.; Harder, U.; Peissner, W.; Wahl, S.; Grallert, H.; Adamski, J.; et al. Changes in the serum metabolite profile in obese children with weight loss. Eur. J. Nutr. 2015,54, 173–181. [CrossRef] 16. Hellmuth, C.; Kirchberg, F.F.; Lass, N.; Harder, U.; Peissner, W.; Koletzko, B.; Reinehr, T. Tyrosine Is Associated with Insulin Resistance in Longitudinal Metabolomic Profiling of Obese Children. J. Diabetes Res. 2016,2016, 2108909. [CrossRef] 17. Hellmuth, C.; Kirchberg, F.F.; Brandt, S.; Moß, A.; Walter, V.; Rothenbacher, D.; Brenner, H.; Grote, V.; Gruszfeld, D.; Socha, P.; et al. An individual participant data meta-analysis on metabolomics profiles for obesity and insulin resistance in European children. Sci. Rep. 2019,9, 5053. [CrossRef] 18. Singh, A.; Kinnebrew, G.; Hsu, P.-C.; Weng, D.Y.; Song, M.-A.; Reisinger, S.A.; McElroy, J.P.; Keller-Hamilton, B.; Ferketich, A.K.; Freudenheim, J.L.; et al. Untargeted Metabolomics and Body Mass in Adolescents: A Cross-Sectional and Longitudinal Analysis. Metabolites 2023,13, 899. [CrossRef] 19. Mansell, T.; Magnussen, C.G.; Nuotio, J.; Laitinen, T.T.; Harcourt, B.E.; Bekkering, S.; McCallum, Z.; Kao, K.-T.; Sabin, M.A.; Juonala, M.; et al. Decreasing severity of obesity from early to late adolescence and young adulthood associates with longitudinal metabolomic changes implicated in lower cardiometabolic disease risk. Int. J. Obes. 2022,46, 646–654. [CrossRef] 20. Ojanen, X.; Cheng, R.; Törmäkangas, T.; Rappaport, N.; Wilmanski, T.; Wu, N.; Fung, E.; Nedelec, R.; Sebert, S.; Vlachopoulos, D.; et al. Towards early risk biomarkers: Serum metabolic signature in childhood predicts cardio-metabolic risk in adulthood. EBioMedicine 2021,72, 1–10. [CrossRef] 21. Hosking, J.; Pinkney, J.; Jeffery, A.; Cominetti, O.; Da Silva, L.; Collino, S.; Kussmann, M.; Hager, J.; Martin, F. Insulin Resistance during normal child growth and development is associated with a distinct blood metabolic phenotype (Earlybird 72). Pediatr. Diabetes 2019,20, 832–841. [CrossRef] 22. Monnerie, S.; Comte, B.; Ziegler, D.; Morais, J.A.; Pujos-Guillot, E.; Gaudreau, P. Metabolomic and Lipidomic Signatures of Metabolic Syndrome and its Physiological Components in Adults: A Systematic Review. Sci. Rep. 2020,10, 669. [CrossRef] 23. Aguilera, C.M.; Gil-Campos, M.; Cañete, R.; Gil, A. Alterations in plasma and tissue lipids associated with obesity and metabolic syndrome. Clin. Sci. 2008,114, 183–193. [CrossRef] [PubMed] 24. Ritter, O.; Jelenik, T.; Roden, M. Lipid-mediated muscle insulin resistance: Different fat, different pathways? J. Mol. Med. Berl. Ger. 2015,93, 831–843. [CrossRef] 25. Kelly, B.; Pearce, E.L. Amino Assets: How Amino Acids Support Immunity. Cell Metab. 2020,32, 154–175. [CrossRef] 26. Zhao, X.; Gang, X.; Liu, Y.; Sun, C.; Han, Q.; Wang, G. Using Metabolomic Profiles as Biomarkers for Insulin Resistance in Childhood Obesity: A Systematic Review. J. Diabetes Res. 2016,2016, 8160545. [CrossRef] 27. Handakas, E.; Keski-Rahkonen, P.; Chatzi, L.; Alfano, R.; Roumeliotaki, T.; Plusquin, M.; Maitre, L.; Richiardi, L.; Brescianini, S.; Scalbert, A.; et al. Cord blood metabolic signatures predictive of childhood overweight and rapid growth. Int. J. Obes. 2021, 45, 2252–2260. [CrossRef] 28. Würtz, P.; Mäkinen, V.P.; Soininen, P.; Kangas, A.J.; Tukiainen, T.; Kettunen, J.; Savolainen, M.J.; Tammelin, T.; Viikari, J.S.; Rönnemaa, T.; et al. Metabolic Signatures of Insulin Resistance in 7,098 Young Adults. Diabetes 2012,61, 1372–1380. [CrossRef] 29. Mohorko, N.; Petelin, A.; Jurdana, M.; Biolo, G.; Jenko-Pražnikar, Z. Elevated serum levels of cysteine and tyrosine: Early biomarkers in asymptomatic adults at increased risk of developing metabolic syndrome. BioMed Res. Int. 2015,2015, 418681. [CrossRef] 30. Zhao, X.; Han, Q.; Liu, Y.; Sun, C.; Gang, X.; Wang, G. The Relationship between Branched-Chain Amino Acid Related Metabolomic Signature and Insulin Resistance: A Systematic Review. J. Diabetes Res. 2016,2016, 2794591. [CrossRef] 31. Guasch-Ferré, M.; Hruby, A.; Toledo, E.; Clish, C.B.; Martínez-González, M.A.; Salas-Salvadó, J.; Hu, F.B. Metabolomics in Prediabetes and Diabetes: A Systematic Review and Meta-analysis. Diabetes Care. 2016,39, 833–846. [CrossRef] 32. Adams, S.H. Emerging perspectives on essential amino acid metabolism in obesity and the insulin-resistant state. Adv. Nutr. Bethesda Md. 2011,2, 445–456. [CrossRef] 33. Walldius, G.; Jungner, I.; Aastveit, A.H.; Holme, I.; Furberg, C.D.; Sniderman, A.D. The apoB/apoA-I ratio is better than the cholesterol ratios to estimate the balance between plasma proatherogenic and antiatherogenic lipoproteins and to predict coronary risk. Clin. Chem. Lab. Med. 2004,42, 1355–1363. [CrossRef] Nutrients 2025,17, 1833 18 of 18 34. Sellers, E.A.C.; Singh, G.R.; Sayers, S.M. Apo-B/AI ratio identifies cardiovascular risk in childhood: The Australian Aboriginal Birth Cohort study. Diabetes Vasc. Dis. Res. 2009,6, 94–99. [CrossRef] 35. Juonala, M.; Viikari, J.S.; Kähönen, M.; Solakivi, T.; Helenius, H.; Jula, A.; Marniemi, J.; Taittonen, L.; Laitinen, T.; Nikkari, T.; et al. Childhood levels of serum apolipoproteins B and A-I predict carotid intima-media thickness and brachial endothelial function in adulthood: The cardiovascular risk in young Finns study. J. Am. Coll. Cardiol. 2008,52, 293–299. [CrossRef] 36. Piperi, C.; Kalofoutis, C.; Papaevaggeliou, D.; Papapanagiotou, A.; Lekakis, J.; Kalofoutis, A. The significance of serum HDL phospholipid levels in angiographically defined coronary artery disease. Clin. Biochem. 2004,37, 377–381. [CrossRef] 37. Fischer, K.; Kettunen, J.; Würtz, P.; Haller, T.; Havulinna, A.S.; Kangas, A.J.; Soininen, P.; Esko, T.; Tammesoo, M.-L.; Mägi, R.; et al. Biomarker profiling by nuclear magnetic resonance spectroscopy for the prediction of all-cause mortality: An observational study of 17,345 persons. PLoS Med. 2014,11, e1001606. [CrossRef] 38. Lawler, P.R.; Akinkuolie, A.O.; Chandler, P.D.; Moorthy, M.V.; Vandenburgh, M.J.; Schaumberg, D.A.; Lee, I.-M.; Glynn, R.J.; Ridker, P.M.; Buring, J.E.; et al. Circulating N-Linked Glycoprotein Acetyls and Longitudinal Mortality Risk. Circ. Res. 2016, 118, 1106–1115. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.