Association between attention-deficit/hyperactivity symptoms and body composition in children: a health examination-based study Authors: Shijie Li, Muhan Li, Mei Xue, Muqiao Yan, Zhuang Wei, Aimin Liang Published: 2026 ABSTRACT BackgroundAlthough attention deficit/hyperactivity disorder (ADHD) has been associated with obesity, research examining specific body composition components and dimensional symptom domains remains limited. This study aimed to investigate the associations between dimensional ADHD symptom domains and specific body composition indicators in children, and to explore potential moderating and mediating pathways.MethodsA retrospective cross-sectional analysis was conducted in 636 children aged 6–18 years. ADHD symptoms were assessed using the parent-reported SNAP-IV scale, and body composition (visceral fat area [VFA]; skeletal muscle mass to weight ratio [SMM/WT]) was assessed via bioelectrical impedance analysis. Multiple linear regression examined independent associations, adjusting for age, sex, and BMI z-score. Moderation and mediation analyses tested whether BMI z-score moderated the SMM/WT-symptom link, and whether metabolic markers mediated the SMM/WT-inattention association.ResultsHigher SMM/WT was independently associated with lower inattention (β = -0.017, p = .041) and hyperactivity-impulsivity scores (β = -0.018, p = .031). Greater VFA was linked to lower hyperactivity-impulsivity (β = -0.002, p = .048) and oppositional defiant symptoms (β = -0.003, p = .011). BMI z-score showed no independent association with any symptom domain and did not moderate the SMM/WT-symptom relationship (all interaction p > .05). None of the metabolic markers (HDL-C, ALT, AST, fasting glucose) significantly mediated the SMM/WT-inattention association (all indirect effect 95% CIs included zero).ConclusionsSpecific body composition components show distinct, symptom-domain-specific associations in children with ADHD, independent of overall weight status. The inverse association of skeletal muscle mass with inattention was not explained by the metabolic markers examined. Background Attention deficit/hyperactivity disorder (ADHD) is a prevalent neurodevelopmental disorder characterized by inattention, hyperactivity, restlessness and impaired impulse control. It represents a major global challenge in pediatric mental health, ranking as the 6th leading cause of years lived with disability (YLDs) among children aged 5–14 years [1, 2]. ADHD exhibits a high prevalence among children and adolescents, affecting approximately 8.0% globally, with boys nearly twice as likely to be affected as girls [3, 4], leading to significant implications for long-term academic, social and psychological outcomes [5]. Beyond its core diagnostic symptoms, ADHD is increasingly recognized as a systemic condition that frequently co-occurs with a broad spectrum of physical and mental health phenotypes, as demonstrated in an umbrella review encompassing over 234 million participants [6]. Among these co-occurring conditions, the link between ADHD and obesity has garnered substantial research interest in recent decades [7–9]. Specifically, children and adolescents with ADHD exhibit a 56% increased likelihood of overweight or obesity compared to neurotypical peers [10], and evidence suggests an elevated risk for early metabolic syndrome and related complications in this population [7, 8]. Notably, subclinical ADHD symptoms, which do not meet full diagnostic criteria but remain functionally impairing, are common, affecting at least a further 5% of children and adolescents [4], and are increasingly linked to adverse lifestyle behaviors and early metabolic risks [10, 11], thereby underscoring the public health importance of investigating this association across the full spectrum of symptom expression. Building on this foundation, research has begun to shift from merely establishing an ADHD-obesity correlation to elucidating the associations with specific aspects of body composition. For instance, emerging evidence, including our prior work, suggests that visceral adiposity, a metabolically active fat depot, may be particularly relevant in children with ADHD and co-occurring obesity [12]. Concurrently, a growing body of literature highlights the potential protective role of skeletal muscle mass against cognitive deficits and neurodevelopmental challenges through mechanisms involving myokine secretion and improved metabolic regulation [13–15]. However, a critical limitation persists in much of the existing literature: ADHD is predominantly treated as a categorical entity. This categorical approach may obscure important, symptom-specific relationships, given that core ADHD dimensions possess distinct neurobiological and behavioral underpinnings [16–19]. In contrast, a dimensional approach conceptualizes ADHD symptoms as continuously distributed traits across the population, rather than a binary categorical diagnosis [20, 21]. Accordingly, the present study investigates symptom-specific correlates, acknowledging that core ADHD dimensions may exhibit heterogeneous associations with metrics such as visceral adiposity and skeletal muscle mass. To address this gap, we conducted a retrospective analysis within a cohort of children undergoing routine health examinations. This study aims to dissect the associations between dimensional ADHD symptom scores and body composition, with a specific focus on visceral fat area and skeletal muscle mass relative to body weight. By adopting this nuanced, symptom-domain-specific approach, our study seeks to provide a more refined understanding of the associations between body composition and ADHD symptomatology. Methods Study design and participants This study is prepared and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement and the Reporting of studies Conducted using Observational Routinely-collected health Data (RECORD) guidelines. This retrospective, cross-sectional study analyzed the medical records of children and adolescents aged 6 to 18 years who underwent a routine, non-referred annual health examination at the Health Care Center of Beijing Children’s Hospital between October 2023 and October 2025. All participants were drawn from a community-based preventive care cohort; none were referred for ADHD evaluation or psychiatric consultation. The initial inclusion criterion was the availability of complete data for body composition analysis, the Swanson, Nolan, and Pelham Rating Scale-IV (SNAP-IV) parent-rated scale [22, 23], and basic demographic information. We applied the following exclusion criteria to define our final analytical sample: (1) underweight (body mass index [BMI] z-score ≤ -2 standard deviations [SD] based on the World Health Organization growth standards) [24], as underweight in children is frequently associated with underlying chronic diseases, malnutrition, or eating disorders that independently alter body composition and may compromise neurobehavioral functioning [25, 26]; (2) a known diagnosis of any neurodevelopmental or psychiatric condition; (3) presence of chronic medical conditions affecting growth, metabolism, or neurodevelopment; (4) current or historical use of psychotropic medications; and (5) missing blood biochemical parameters. The study protocol was approved by the Institutional Review Board of Beijing Children’s Hospital (2025-027-D), and the requirement for informed consent was waived due to the retrospective nature of the study. Clinical trial number: not applicable. ADHD symptom dimensions The severity of ADHD and oppositional defiant disorder (ODD) symptoms was assessed using the Chinese parent-report version of the Swanson, Nolan, and Pelham Rating Scale-IV (SNAP-IV) [22, 23], administered as part of the standard developmental screening protocol during routine health checkups. This version has been previously validated in the Chinese population and exhibits robust psychometric properties [27, 28]. The scale comprises three core subscales: Inattention (9 items), Hyperactivity-Impulsivity (9 items), and Oppositional Defiant (8 items). Parents rated the frequency of each behavior over the preceding month on a 4-point Likert scale, ranging from 0 (not at all) to 3 (very much). The average score for each subscale was calculated and used as a continuous variable in all statistical analyses. Body composition analysis Body composition was assessed using a direct segmental multi-frequency bioelectrical impedance analysis (BIA) device (InBody 770, InBody Co., Ltd., Seoul, Korea). All assessments were conducted by trained technicians following standardized protocols in the morning, with participants wearing light clothing and having fasted for at least 8 h. Key parameters extracted for analysis included the visceral fat area (VFA) and the skeletal muscle mass to weight ratio (SMM/WT). Blood biochemical analysis Fasting venous blood samples were collected from all participants in the morning after an overnight fast of at least 8 h. Blood was drawn into serum separator tubes and centrifuged within 30 min of collection to minimize ex vivo glycolysis. Serum was analyzed using standard automated clinical chemistry analyzers for the following metabolic markers: liver enzymes (alanine aminotransferase [ALT] and aspartate aminotransferase [AST]), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C) and fasting blood glucose (GLU). Statistical analysis Statistical analyses were performed using SPSS version 25.0 and the PROCESS macro for SPSS (version 4.1). Descriptive statistics are presented as means ± standard deviations, medians with interquartile ranges, or counts with percentages, as appropriate. Spearman’s rank correlation was used to examine bivariate associations among key demographic, body composition, metabolic, and symptom variables. To examine the independent associations between body composition and ADHD symptom domains, we fitted separate multiple linear regression models for each SNAP-IV subscale score (Inattention, Hyperactivity-Impulsivity, Oppositional Defiant). Only VFA and SMM/WT were retained as primary body composition predictors; total body fat mass (BFM) was excluded because preliminary diagnostics revealed severe multicollinearity with BMI z-score (VIF = 30.55) and VFA (VIF = 27.95). VFA was retained because visceral adipose tissue is metabolically active and specifically implicated in inflammatory and neurodevelopmental pathways [29]. All models included the following covariates: age, sex, BMI z-score, VFA, and SMM/WT. The covariates were chosen a priori based on their potential confounding influence. All variance inflation factor (VIF) values were below 5, indicating that multicollinearity did not unduly influence the stability of the regression estimates. We further conducted moderation and mediation analyses to elucidate the nature of the observed associations. Moderation analysis (using PROCESS Model 1) tested whether the association between SMM/WT and ADHD symptoms was moderated by the child’s overall weight status (BMI z-score). Mediation analysis (using PROCESS Model 4) tested whether significant associations between SMM/WT and symptom domains were mediated by metabolic markers (HDL-C, ALT, AST, GLU). For all mediation analyses, bias-corrected bootstrap confidence intervals were based on 5000 samples, and effects were considered significant if the 95% confidence interval did not contain zero. All tests were two-tailed, and a p-value < 0.05 was considered statistically significant. Results Characteristics of the study population The initial screening of medical records identified a pool of 874 children and adolescents who had undergone the standard health examination protocol. After applying the pre-defined exclusion criteria, the final analytic cohort consisted of 636 non-underweight children. Exclusions were made for the following reasons: underweight status (n = 36), neurodevelopmental or psychiatric diagnoses (n = 11), psychotropic medications use (n = 12; including current or historical use of methylphenidate, atomoxetine, sertraline, risperidone, or aripiprazole), chronic medical conditions affecting growth or metabolism (n = 35; including endocrine, respiratory, cardiovascular, renal, and hematological disorders requiring ongoing treatment), and missing blood biochemical parameters (n = 144). Demographic characteristics and ADHD symptom profiles are summarized in Table 1. The cohort had a median age of 9.8 years (8.0, 12.5), with males accounting for 64.5% of the sample. The median BMI z-score was 0.36 (-0.59, 1.66), indicating that the majority of participants were within the normal weight range. Of note, 119 children (18.7%) had a BMI z-score > 2 SD, consistent with obesity. Parent-reported ADHD symptom scores on the SNAP-IV scale showed a mean inattention score of 0.89 ± 0.50 (range 0.11–2.78), a median hyperactivity-impulsivity score of 0.56 (0.22, 0.49; range 0.00–2.78), and a mean oppositional defiant score of 0.88 ± 0.49 (range 0.00–2.75). Table 1Characteristics of the study participants (N = 636)CharacteristicValueDemographics Age, years9.8 (8.0, 12.5) Sex, boys, n (%)64.5 BMI z-score0.36 (-0.59, 1.66)SNAP-IV Symptom Scores Inattention0.89 ± 0.50 Hyperactivity-Impulsivity0.56 (0.22, 1.00) Oppositional Defiant0.88 ± 0.49Data for continuous variables are presented as Mean ± Standard Deviation for normally distributed variables and as Median [Interquartile Range] for non-normally distributed variables. Categorical data are presented as n (%) Table 1Characteristics of the study participants (N = 636)CharacteristicValueDemographics Age, years9.8 (8.0, 12.5) Sex, boys, n (%)64.5 BMI z-score0.36 (-0.59, 1.66)SNAP-IV Symptom Scores Inattention0.89 ± 0.50 Hyperactivity-Impulsivity0.56 (0.22, 1.00) Oppositional Defiant0.88 ± 0.49Data for continuous variables are presented as Mean ± Standard Deviation for normally distributed variables and as Median [Interquartile Range] for non-normally distributed variables. Categorical data are presented as n (%) Characteristics of the study participants (N = 636) Data for continuous variables are presented as Mean ± Standard Deviation for normally distributed variables and as Median [Interquartile Range] for non-normally distributed variables. Categorical data are presented as n (%) Bivariate correlations among key variables Spearman correlation analysis revealed several significant bivariate associations (Table 2). Female sex was negatively correlated with scores for inattention (IA, ρ = -0.12, p < .01), hyperactivity-impulsivity (HI, ρ = -0.22, p < .01), and oppositional defiant (ODD, ρ = -0.10, p < .05) symptoms. Older age was negatively correlated with HI scores (ρ = -0.31, p < .01). A higher BMI z-score was positively correlated with IA (ρ = 0.12, p < .01) and HI (ρ = 0.10, p < .05) scores. A higher SMM/WT was negatively correlated with IA (ρ = -0.10, p < .05) and HI (ρ = -0.09, p < .05) scores. VFA was negatively correlated with HI scores (ρ = -0.13, p < .01). Among the body composition indicators, BMI z-score was positively correlated with VFA (ρ = 0.73, p < .01) and negatively correlated with SMM/WT (ρ = -0.67, p < .01). VFA and SMM/WT were negatively correlated (ρ = -0.65, p < .01). Table 2Multiple linear regression analyses of ADHD symptom dimensions on body composition indicatorsBivariate12345678910111213141.Gender-2.Age-0.01-3.BMIZ− 0.26**0.15**-4.VFA-0.010.57**0.73**-5.SMMWT− 0.12**0.08*− 0.67**− 0.65**-6.IA− 0.12**-0.050.12**0.03− 0.10*-7.HI− 0.22**− 0.31**0.10*− 0.13**− 0.09*0.65**8.ODD− 0.10*-0.040.04-0.04-0.010.55**0.54**-9.TC0.07-0.040.13**0.11**− 0.21**0.030.030-10.GLU− 0.16**0.050.21**0.13**-0.040.030.040.01-0.01-11.AST− 0.13**− 0.49**-0.06− 0.30**− 0.08*0.060.17**0.010.10*-0.07-12.ALT− 0.22**0.040.41**0.30**− 0.33**0.11**0.11**0.060.12**0.070.50**-13.TG0.030.33**0.35**0.45**− 0.27**0-0.07-0.010.22**0.07− 0.14**0.20**-14.HDLC0.01− 0.22**− 0.32**− 0.38**0.25**-0.020.060.020.27**-0.060.16**− 0.15**− 0.37**-*p < .05, **p < .01. Gender: Female = 1, Male = 0 Table 2Multiple linear regression analyses of ADHD symptom dimensions on body composition indicatorsBivariate12345678910111213141.Gender-2.Age-0.01-3.BMIZ− 0.26**0.15**-4.VFA-0.010.57**0.73**-5.SMMWT− 0.12**0.08*− 0.67**− 0.65**-6.IA− 0.12**-0.050.12**0.03− 0.10*-7.HI− 0.22**− 0.31**0.10*− 0.13**− 0.09*0.65**8.ODD− 0.10*-0.040.04-0.04-0.010.55**0.54**-9.TC0.07-0.040.13**0.11**− 0.21**0.030.030-10.GLU− 0.16**0.050.21**0.13**-0.040.030.040.01-0.01-11.AST− 0.13**− 0.49**-0.06− 0.30**− 0.08*0.060.17**0.010.10*-0.07-12.ALT− 0.22**0.040.41**0.30**− 0.33**0.11**0.11**0.060.12**0.070.50**-13.TG0.030.33**0.35**0.45**− 0.27**0-0.07-0.010.22**0.07− 0.14**0.20**-14.HDLC0.01− 0.22**− 0.32**− 0.38**0.25**-0.020.060.020.27**-0.060.16**− 0.15**− 0.37**-*p < .05, **p < .01. Gender: Female = 1, Male = 0 Multiple linear regression analyses of ADHD symptom dimensions on body composition indicators *p < .05, **p < .01. Gender: Female = 1, Male = 0 Independent associations of body composition with ADHD symptom domains Multiple linear regression analyses, adjusted for age, sex, and BMI z-score, revealed distinct associations between body composition indicators and ADHD symptom domains (Table 3). Table 3 Examination of metabolic markers as potential mediators between skeletal muscle mass and inattentive symptomsPredictorInattentionHyperactivity-ImpulsivityOppositional Defiantβ (SE)95% CI P β (SE)95% CI P β (SE)95% CI P Sex (Female)-0.140 (0.047)[-0.232, -0.047]0.003-0.265 (0.047)[-0.358, -0.172]< 0.001-0.091 (0.047)[-0.183, 0.001]0.052Age (months)0.000 (0.001)[-0.002, 0.002]0.803-0.003 (0.001)[-0.005, -0.002]< 0.0010.002 (0.001)[0.000, 0.003]0.071BMI z-score0.013 (0.024)[-0.033, 0.060]0.5810.025 (0.024)[-0.022, 0.072]0.2960.027 (0.024)[-0.020, 0.073]0.261VFA-0.001 (0.001)[-0.003, 0.001]0.226-0.002 (0.001)[-0.004, 0.000]0.048-0.003 (0.001)[-0.005, -0.001]0.011SMM/WT-0.017 (0.008)[-0.034, -0.001]0.041-0.018 (0.008)[-0.035, -0.002]0.031-0.015 (0.008)[-0.031, 0.002]0.080Adjusted R²0.0230.1570.009All models are adjusted for all other variables in the tableβ unstandardized regression coefficient, SE standard error Table 3 Examination of metabolic markers as potential mediators between skeletal muscle mass and inattentive symptomsPredictorInattentionHyperactivity-ImpulsivityOppositional Defiantβ (SE)95% CI P β (SE)95% CI P β (SE)95% CI P Sex (Female)-0.140 (0.047)[-0.232, -0.047]0.003-0.265 (0.047)[-0.358, -0.172]< 0.001-0.091 (0.047)[-0.183, 0.001]0.052Age (months)0.000 (0.001)[-0.002, 0.002]0.803-0.003 (0.001)[-0.005, -0.002]< 0.0010.002 (0.001)[0.000, 0.003]0.071BMI z-score0.013 (0.024)[-0.033, 0.060]0.5810.025 (0.024)[-0.022, 0.072]0.2960.027 (0.024)[-0.020, 0.073]0.261VFA-0.001 (0.001)[-0.003, 0.001]0.226-0.002 (0.001)[-0.004, 0.000]0.048-0.003 (0.001)[-0.005, -0.001]0.011SMM/WT-0.017 (0.008)[-0.034, -0.001]0.041-0.018 (0.008)[-0.035, -0.002]0.031-0.015 (0.008)[-0.031, 0.002]0.080Adjusted R²0.0230.1570.009All models are adjusted for all other variables in the tableβ unstandardized regression coefficient, SE standard error Examination of metabolic markers as potential mediators between skeletal muscle mass and inattentive symptoms All models are adjusted for all other variables in the table β unstandardized regression coefficient, SE standard error For inattention, a higher SMM/WT was associated with lower scores (β = -0.017, p= .041). Female sex was also associated with lower inattention scores (β = -0.140, p= .003). VFA, age, and BMI z-score were not significantly associated. For hyperactivity-impulsivity, a higher SMM/WT was associated with lower scores (β = -0.018, p= .031). A larger VFA was also independently associated with lower scores (β = -0.002, p = .048). Female sex (β = -0.265, p< .001) and older age (β = -0.003, p< .001) were also associated with lower scores. BMI z-score was not a significant predictor. For oppositional defiant symptoms, a larger VFA was associated with lower scores (β = -0.003, p= .011). The association for SMM/WT was not significant (β = -0.015, p= .080). The associations for sex and age were not significant. BMI z-score was not independently associated with any of the three symptom domains in the multivariate models (all p > .05). Absence of moderating effect by weight status We further tested whether the protective association between the SMM/WT and ADHD symptoms was moderated by the child’s overall weight status (BMI z-score). This was examined by including an interaction term (SMM/WT × BMI z-score) in the multiple linear regression models for each symptom domain. The interaction term was not statistically significant for inattention, hyperactivity-impulsivity, or oppositional defiant symptoms (all p > .05). This indicates that the strength of the association between SMM/WT and ADHD symptom severity did not vary significantly across different levels of BMI z-score (Fig. 1). Fig. 1Association between skeletal muscle mass and ADHD symptoms across BMI z-score levels. Scatter plots with regression lines show the relationship between skeletal muscle mass to weight ratio (SMM/WT) and scores for Inattention, Hyperactivity-Impulsivity, and Oppositional Defiant symptoms at different levels of BMI z-score. The absence of a significant moderating effect indicates that the association between SMM/WT and ADHD symptoms did not vary with overall weight status Association between skeletal muscle mass and ADHD symptoms across BMI z-score levels. Scatter plots with regression lines show the relationship between skeletal muscle mass to weight ratio (SMM/WT) and scores for Inattention, Hyperactivity-Impulsivity, and Oppositional Defiant symptoms at different levels of BMI z-score. The absence of a significant moderating effect indicates that the association between SMM/WT and ADHD symptoms did not vary with overall weight status Examination of potential metabolic mediators Mediation analyses were conducted to test if the association between SMM/WT and inattention was mediated by metabolic markers (HDL-C, AST, ALT, Fasting Glucose). LDL-C and TG were measured but excluded from primary mediation analyses because they showed no significant bivariate association with inattention symptoms, violating the statistical prerequisite for testing mediation. The results are presented in Table 4. Table 4Examination of metabolic markers as potential mediators between skeletal muscle mass and inattentive symptomsPotential MediatorPath a: SMM/WT→Mediator (β, p)Path b: Mediator→IA (β, p)Indirect Effect (Bootstrapped 95% CI)Direct Effect (SMM/WT→IA) (β, p)HDL-C0.0163, < 0.0010.0366, 0.5460.0006 (-0.0013, 0.0028)-0.0110, 0.0095AST-0.3112, 0.001-0.0007, 0.9370.0002 (-0.0026, 0.0024)-0.0110, 0.0095ALT-1.3648, < 0.0010.0007, 0.615-0.0010 (-0.0054, 0.0030)-0.0110, 0.0095Fasting Glucose-0.0054, 0.0950.0339, 0.214-0.0002 (-0.0010, 0.0004)-0.0110, 0.0095The direct effect represents the effect of SMM/WT on IA after controlling for the mediator and covariates. Path a represents the effect of the independent variable (SMM/WT) on the mediator. Path b represents the effect of the mediator on the dependent variable (IA), controlling for SMM/WT. All models are adjusted for age, sex, and BMI z-score. The statistical significance of the indirect effect is determined by a bootstrapped 95% confidence interval (CI) that does not contain zeroSMM/WT Skeletal muscle mass to weight ratio, IA Inattention, HDL-C High-density lipoprotein cholesterol, AST Aspartate aminotransferase, ALT Alanine aminotransferase Table 4Examination of metabolic markers as potential mediators between skeletal muscle mass and inattentive symptomsPotential MediatorPath a: SMM/WT→Mediator (β, p)Path b: Mediator→IA (β, p)Indirect Effect (Bootstrapped 95% CI)Direct Effect (SMM/WT→IA) (β, p)HDL-C0.0163, < 0.0010.0366, 0.5460.0006 (-0.0013, 0.0028)-0.0110, 0.0095AST-0.3112, 0.001-0.0007, 0.9370.0002 (-0.0026, 0.0024)-0.0110, 0.0095ALT-1.3648, < 0.0010.0007, 0.615-0.0010 (-0.0054, 0.0030)-0.0110, 0.0095Fasting Glucose-0.0054, 0.0950.0339, 0.214-0.0002 (-0.0010, 0.0004)-0.0110, 0.0095The direct effect represents the effect of SMM/WT on IA after controlling for the mediator and covariates. Path a represents the effect of the independent variable (SMM/WT) on the mediator. Path b represents the effect of the mediator on the dependent variable (IA), controlling for SMM/WT. All models are adjusted for age, sex, and BMI z-score. The statistical significance of the indirect effect is determined by a bootstrapped 95% confidence interval (CI) that does not contain zeroSMM/WT Skeletal muscle mass to weight ratio, IA Inattention, HDL-C High-density lipoprotein cholesterol, AST Aspartate aminotransferase, ALT Alanine aminotransferase Examination of metabolic markers as potential mediators between skeletal muscle mass and inattentive symptoms The direct effect represents the effect of SMM/WT on IA after controlling for the mediator and covariates. Path a represents the effect of the independent variable (SMM/WT) on the mediator. Path b represents the effect of the mediator on the dependent variable (IA), controlling for SMM/WT. All models are adjusted for age, sex, and BMI z-score. The statistical significance of the indirect effect is determined by a bootstrapped 95% confidence interval (CI) that does not contain zero SMM/WT Skeletal muscle mass to weight ratio, IA Inattention, HDL-C High-density lipoprotein cholesterol, AST Aspartate aminotransferase, ALT Alanine aminotransferase A higher SMM/WT was significantly associated with higher HDL-C (β = 0.0163, p< .001) and lower ALT (β = -1.3648, p< .001) and AST (β = -0.3112, p= .001) levels. It was not significantly associated with fasting glucose (β = -0.0054, p= .095). None of the metabolic markers were significantly associated with inattention symptoms after controlling for SMM/WT and covariates (HDL-C: β = 0.0366, p= .546; ALT: β = 0.0007, p= .615; AST: β = -0.0007, p= .937; Fasting Glucose: β = 0.0339, p= .214). The bootstrapped 95% confidence intervals for all indirect effects included zero (HDL-C: -0.0013 to 0.0028; ALT: -0.0054 to 0.0030; AST: -0.0026 to 0.0024; Fasting Glucose: -0.0010 to 0.0004). The direct effect of SMM/WT on inattention remained significant after including the mediators (β = -0.0110, p= .0095). Discussion This study provides evidence for distinct associations between specific body composition metrics and dimensional ADHD symptomatology in a large sample of children and adolescents undergoing routine health examinations. After adjusting for key potential confounders including age, sex, and overall adiposity (BMI z-score), our analyses revealed that higher relative muscle mass is independently linked to lower severity of both inattention and hyperactivity-impulsivity, whereas greater visceral adiposity is associated with reduced hyperactivity-impulsivity and oppositional defiant symptoms. These findings underscore the value of moving beyond aggregate weight metrics to decipher the distinct phenotypic correlates of ADHD symptomatology. Our results reveal a symptom-domain-specific relationship between body composition and ADHD. Although inattention and hyperactivity-impulsivity frequently co-occur, they demonstrate distinct developmental trajectories and correlates, as reflected by their differential associations with age and sex in our study and previous reports, where hyperactivity-impulsivity typically declines with age and is more common in males [30, 31], whereas inattention shows a less marked age-related decrease and a weaker association with sex [32, 33]. Within this context, higher skeletal muscle mass was most consistently associated with lower inattention, with a similar but less pronounced association observed for hyperactivity-impulsivity. This negative association is biologically plausible given hypothesized roles of myokines such as brain-derived neurotrophic factor (BDNF) and irisin in supporting neurotrophic signaling and neuroplasticity within fronto-striatal and fronto-parietal networks [34]. Direct evidence linking specific myokines to ADHD symptom variation in children is currently lacking. Alternatively, from a behavioral perspective, children with prominent inattention often present with impairments in executive function, motor planning, and behavioral organization [32], which may hinder their consistent participation in structured physical activities or regular exercise routines. Over time, such reduced physical engagement could manifest as lower accumulation of relative skeletal muscle mass compared to peers with fewer inattentive symptoms. Thus, the observed muscle-inattention association may partly reflect behavioral patterns associated with the symptom domain, rather than representing a direct muscle-brain molecular pathway. In contrast, greater visceral adiposity was associated with lower hyperactivity-impulsivity and oppositional defiant symptoms. This unexpected finding should not be interpreted as evidence that visceral fat is protective. Unlike the association between skeletal muscle and ADHD symptoms, where exercise intervention trials provide supportive evidence for domain-specific behavioral improvements [35], no comparable evidence suggests that increasing visceral fat benefits behavioral regulation. We interpret this finding cautiously. It may reflect the distinct metabolic context of our predominantly normal-weight sample, where adequate energy reserves within the physiological range may support behavioral stability by ensuring baseline metabolic sufficiency. Adipose tissue functions as an active endocrine organ secreting hormones that regulate energy homeostasis, and these functions may remain operative within a healthy metabolic range [36]. Alternatively, this negative association may partly reflect reverse causation, whereby children with higher hyperactivity-impulsivity expend more energy and accumulate less visceral fat. This interpretation is based on behavioral observation rather than longitudinal evidence specifically linking hyperactivity-impulsivity to visceral fat reduction. These interpretations are not mutually exclusive, but all remain hypothetical given our cross-sectional design. The adverse cardiometabolic and inflammatory effects of pathological visceral adiposity in obesity remain undisputed [37, 38]. Critically, these findings do not imply that maintaining adiposity within a normal range could reduce the risk or severity of ADHD. Interventions targeting body composition in this population should be framed as strategies to promote metabolic health, rather than as primary or adjunctive treatments for core neurodevelopmental symptoms. A critical insight from our analysis is that specific body composition components, namely relative muscle mass and visceral adiposity, appear to be more pertinent to ADHD symptomatology than overall body weight or BMI. This clarifies prior research that has inconsistently linked ADHD to obesity as a categorical entity, with some studies reporting positive associations [37, 38] and others null findings [39]. Our results raise the possibility that this heterogeneity partly reflects BMI’s limited sensitivity as an aggregate metric. The metabolic and neurotrophic profiles of specific fat depots and muscle tissue may represent more direct tissue-level correlates, which overall adiposity captures only partially. The absence of an independent BMI effect in our multivariate models, observed in a predominantly normal-weight community sample, is consistent with the view that body composition quality warrants attention beyond crude weight status. We explored whether the association between muscle mass and inattention was mediated by key metabolic markers. The lack of significant mediation via HDL-C, liver enzymes, or fasting glucose suggests that the observed association between higher muscle mass and lower inattention may not operate primarily through these standard systemic metabolic pathways. This directs future research toward alternative mechanistic pathways. Potential candidates include the modulation of chronic, low-grade inflammation, as myokines possess anti-inflammatory properties, while visceral fat is a source of pro-inflammatory cytokines that can affect neuroinflammation [40, 41]. The gut-microbiota-brain axis represents another promising avenue, given its links to both body composition and neurodevelopment [42]. Muscle contraction releases myokines such as irisin [43, 44]. Although the precise mechanisms remain under investigation, irisin is known to cross the blood-brain barrier and has been associated with neurobiological processes that are relevant to ADHD, including synaptic plasticity and neurogenesis [13, 15]. Whether these mechanisms specifically mediate ADHD symptom variation remains to be established. Our negative mediation results thus help refine the hypothetical framework for the muscle-brain connection in ADHD. From a clinical perspective, these findings highlight the potential value of body composition assessment, beyond BMI, in the management of children with elevated ADHD symptoms. The associations observed here suggest that promoting a metabolically healthy phenotype characterized by higher relative muscle mass warrants consideration as a supportive health goal. Concurrently, this perspective aligns with convergent evidence from meta-analyses showing that physical activity interventions, particularly those involving cognitively engaging exercise, can improve core ADHD symptoms, enhance executive functions (e.g., working memory, cognitive flexibility), and promote motor skills in this population [35, 45]. Interventions involving structured, cognitively engaging physical activities may improve ADHD symptoms through multiple pathways, including but not limited to changes in body composition. This shifts the focus from mere weight control to promoting a metabolically healthy phenotype characterized by higher muscle mass. Such approaches could be beneficial even for children within a normal BMI range who may have adverse body composition profiles [46, 47]. However, a crucial caveat must be noted regarding the practical translation of these findings. While evaluating skeletal muscle and fat metrics provides valuable research insights, implementing routine body composition assessments in clinical practice requires significant caution. Individuals with ADHD exhibit substantially elevated rates of co-occurring eating disorders and body image disturbances compared to neurotypical peers [48]. Consequently, clinicians should avoid overemphasizing weight or tissue metrics to prevent inadvertently worsening body dissatisfaction or triggering disordered eating behaviors in this vulnerable pediatric population. Metabolic evaluations must always be balanced with psychological screenings. Previously, we found that children with ADHD and obesity had significantly higher visceral fat content compared to children with obesity but without ADHD [12]. The role of fat in relation to behavioral outcomes may vary across developmental stages and metabolic contexts, alongside factors such as sex and muscle mass. The cross-sectional design precludes causal inference, and bidirectional relationships remain plausible. Children with severe inattention may struggle to adhere to structured physical activities or maintain regular dietary patterns, which could result in lower relative muscle mass. Conversely, children with higher hyperactivity-impulsivity may expend more energy, thereby limiting visceral fat accumulation. These reverse pathways are consistent with longitudinal evidence that ADHD symptoms predict subsequent increases in adiposity [38, 49]. Whether this pathway extends to visceral fat distribution, or whether the inverse association observed in our cross-sectional data reflects a distinct mechanism, remains to be clarified. Prospective and interventional studies are needed to clarify the directionality of these associations. Several limitations of this study must be acknowledged. Its cross-sectional design precludes any causal inference regarding the direction of the observed associations. Important confounding variables, such as dietary patterns, detailed physical activity levels, sleep quality, and screen time, were not available and could influence both body composition and ADHD symptoms [50, 51]. Because our participants entered the study at 6 years of age or older, we were unable to reconstruct individual growth trajectories or determine the timing of adiposity rebound at the individual level. Adiposity rebound typically occurs between 5 and 7 years of age and has been associated with subsequent weight-related outcomes in longitudinal studies [52]. Future prospective studies incorporating repeated anthropometric assessments may help clarify whether adiposity rebound timing influences associations between body composition and ADHD symptom dimensions. Although our sample was community-based, it was drawn from a single tertiary hospital and consisted of children presenting for health examinations, which may select for families with greater health awareness and socioeconomic advantages, potentially limiting the generalizability of our findings. Conclusions This study demonstrates that specific components of body composition, namely skeletal muscle mass and visceral fat, show distinct associations with different ADHD symptom domains, independent of overall body weight. These associations appear to be independent of several key metabolic markers examined in the present study. Our findings advocate for a more nuanced approach to evaluating pediatric metabolic health, emphasizing the clinical value of monitoring body composition quality over mere weight status in children with elevated ADHD symptoms. Future longitudinal and prospective intervention studies are needed to elucidate causality and explore the potential role of body composition optimization in promoting systemic metabolic health and long-term physical well-being in this pediatric population.