Dietary Index for Gut Microbiota and the Odds of Metabolic Dysfunction‐Associated Fatty Liver Disease in Overweight and Obese Children and Adolescents: A Cross‐Sectional Study Authors: Farzam Kamrani, Ali Nikparast, Ali Tabatabaeyan, Elahe Etesami, Pejman Rohani, Golaleh Asghari Published: 2026 ABSTRACT ABSTRACTBackground and AimsMetabolic dysfunction‐associated fatty liver disease (MAFLD) is increasingly prevalent among children and adolescents with overweight or obesity, posing a growing public health challenge. Emerging evidence highlights the gut microbiota as a critical mediator in MAFLD pathogenesis, yet few studies have explored dietary factors influencing this axis in pediatric populations. This cross‐sectional study investigated the association between the Dietary Index for Gut Microbiota (DI‐GM) and MAFLD.MethodsA total of 505 Iranian youths aged 7–18 years enrolled in a national obesity registry. Dietary intake was assessed via a validated 147‐item Food Frequency Questionnaire. MAFLD was diagnosed according to international expert consensus, requiring ultrasound‑confirmed hepatic steatosis in the presence of overweight/obesity or other metabolic abnormalities.ResultsMultivariable logistic regression models revealed that higher DI‐GM scores were significantly associated with lower odds of MAFLD (OR per unit increase: 0.80, 95% CI: 0.71–0.91; p < 0.001). Participants in the highest DI‐GM quartile had markedly reduced risk compared to the lowest quartile (OR: 0.42, 95% CI: 0.22–0.81). Notably, the inverse association was significant only among post‐pubertal participants. Mediation analysis indicated that liver enzymes—alanine aminotransferase (ALT), gamma‐glutamyl transferase (GGT), and aspartate aminotransferase (AST)—partially explained this relationship, accounting for 20.0%, 19.2%, and 6.7% of the association, respectively.ConclusionThese findings underscore the potential of microbiota‐oriented dietary strategies to mitigate MAFLD risk in pediatric populations with obesity.Keywords: adolescent, children, Dietary Index for Gut Microbiota, MAFLD, metabolic dysfunction‐associated fatty liver disease Introduction The rising prevalence of metabolic dysfunction‐associated fatty liver disease (MAFLD) represents an urgent global public health challenge, particularly among pediatric populations with excess body weight, where the reported prevalence reaches 50.7% [1]. MAFLD, defined as hepatic steatosis in the presence of metabolic perturbations, is strongly associated with a spectrum of cardiometabolic complications, including insulin resistance, dyslipidemia, and type 2 diabetes mellitus (T2DM) [2, 3, 4]. Its progressive course can result in advanced hepatic complications, including fibrosis, cirrhosis, and hepatocellular carcinoma, contributing to substantial long‐term morbidity [4]. The pathogenesis of MAFLD is multifactorial, involving intricate interactions among genetic susceptibility, metabolic dysregulation, chronic systemic inflammation, oxidative stress, environmental influences, and alterations in gut microbiota composition and function [5, 6]. Given the lack of approved pharmacological therapies, lifestyle‐based interventions—particularly dietary modification, regular physical activity, and sustained weight loss—remain the cornerstone of both prevention and attenuation of disease progression [5]. The gut–liver axis describes the bidirectional relationship between the gastrointestinal tract, its resident microbiota, and the liver, modulated by a complex interplay of dietary, genetic, and environmental factors [7]. In individuals with MAFLD, this axis is often disrupted, with gut microbial dysbiosis being a prominent feature. Specifically, patients exhibit an increased relative abundance of potentially pathogenic bacterial genera such as Proteus, Enterobacter, and Escherichia, coupled with a decreased presence of beneficial commensals including Ruminococcus and Lactobacillus [8]. Dysbiosis can compromise the integrity of the intestinal barrier, which facilitates the translocation of microbial‐derived products, particularly lipopolysaccharides, into the portal circulation [8, 9]. This translocation is known to promote hepatic inflammation, exacerbate insulin resistance, and contribute to the accumulation of lipids within hepatocytes [8, 9]. Collectively, these mechanisms not only initiate but also accelerate the progression of liver injury. This underscores the importance of the gut microbiota as a potential therapeutic target in the management of MAFLD. In 2024, Kase et al. conducted a literature‐based study to develop a Dietary Index for Gut Microbiota (DI‐GM), identifying dietary components associated with both beneficial and detrimental effects on gut microbiota; DI‐GM was developed by associating dietary items with biomarkers of microbial diversity—urinary enterodiol and enterolactone [10]. Emerging evidence indicates that for each point increase in the DI‐GM score, there is an 8% reduction in the prevalence of nonalcoholic fatty liver disease (NAFLD) [11], a 6.1% decrease in the prevalence of MAFLD [12], and a 2.5% reduction in the risk of metabolic syndrome (MetS) when the DI‐GM score exceeds the threshold of 3 [13] within the American adults. In addition, findings from mediation analyses have revealed that inflammatory markers such as the Systemic Immune‐Inflammation Index (SII) and the neutrophil‐to‐lymphocyte ratio (NLR) mediate the relationship between DI‐GM and MetS [13]. This underscores that inflammation may be triggered by structural components of bacteria, potentially initiating a cascade of inflammatory pathways involving interleukins and other cytokines. Additionally, by‐products of metabolic processes within bacteria, including certain short‐chain fatty acids, may play a role in inhibiting inflammatory processes [14]. To the best of our knowledge, no previous study has investigated the association between DI‐GM and the odds of MAFLD among overweight and obese children and adolescents. Given the unique physiological and developmental characteristics of the pediatric population, as well as the early onset of metabolic derangements, this gap represents a critical area of unmet research need. Therefore, the present study aimed to assess the relationship between DI‐GM and the odds of MAFLD in this vulnerable group. Additionally, we conducted mediation analyses to explore whether cardiometabolic markers may be involved in the potential mechanisms linking diet‐related modulation of gut microbiota with MAFLD odds. These findings may offer novel insights into microbiota‐targeted dietary strategies for the early prevention and management of MAFLD in youth. Methods Study Design and Population This cross‐sectional study was carried out between September 2023 and July 2024 as part of the Obesity Registry Program in children at the Tehran University of Medical Sciences [15]. The dataset has previously contributed to several published investigations exploring cardiometabolic risk factors and dietary patterns among Iranian youth [16, 17, 18]. Participants were randomly selected from patients attending outpatient clinics in gastroenterology, hepatology, endocrinology, and nutrition at the Children's Medical Center, affiliated with the Tehran University of Medical Sciences. Inclusion criteria comprised children and adolescents aged 7–18 years who were classified as overweight or obese, defined by a body mass index‐for‐age (BMI‐for‐age) Z‐score ≥ 1, in accordance with the World Health Organization (WHO) growth reference standards [19]. Exclusion criteria included: (1) known hepatic, renal, thyroid, or oncological conditions (e.g., Wilson's disease, autoimmune hepatitis, hemochromatosis, viral hepatitis, and malignancies); (2) use of medications known to induce hepatic steatosis or liver injury, including valproic acid and amiodarone, as well as anti‐obesity or appetite‐suppressing agents; (3) major intentional dietary modifications within the past year, including medically prescribed therapeutic diets (e.g., low‑calorie, ketogenic, elimination, or weight‑loss diets), or substantial dietary restriction initiated due to illness or structured weight‑management programs; and (4) incomplete dietary data, including fewer than 35 completed items on the Food Frequency Questionnaire (FFQ), or implausible total energy intake reports. To identify under‐ and over‐reporters, individual energy intake was compared against estimated energy requirements, with exclusions made for deviations exceeding ±2 standard deviation (SD), in line with Institute of Medicine (IOM) recommendations [20]. It is important to note that high reported intake may reflect truly high consumption rather than over‐reporting; thus, the exclusion primarily aims to minimize bias introduced by implausible reports that are likely due to misreporting or recall errors. The study protocol was reviewed and approved by the Ethics Committee of the National Nutrition and Food Technology Research Institute (IR.SBMU.NNFTRI.REC.1402.015) and adhered to the ethical principles set forth in the Declaration of Helsinki. Written informed consent was obtained from the participants' parents or legal guardians, and assent was secured from all children and adolescents enrolled in the study. Anthropometric and Clinical Assessments Anthropometric measurements were performed by trained pediatric nutritionists in accordance with standardized procedures. Body weight was measured to the nearest 100 g using a calibrated digital scale (Seca, Hamburg, Germany), while height was assessed to the nearest 0.5 cm using a fixed stadiometer. Body mass index (BMI) was subsequently calculated as weight (kg) divided by height squared (m2), and BMI‐for‐age Z‐scores were derived using WHO growth standards. Based on WHO criteria, overweight and obesity were defined as BMI‑for‑age Z‑scores > +1 SD and > +2 SD, respectively [19]. Waist circumference (WC) was measured at the midpoint between the lowest rib and the iliac crest, using a nonstretchable measuring tape with 0.5 cm precision. Pubertal development was evaluated by a pediatric endocrinologist using the Marshall–Tanner staging system, which classifies participants as prepubertal or pubertal based on breast development in girls and genital development in boys [21, 22]. Physical activity levels were assessed using the Persian version of the Modifiable Activity Questionnaire (MAQ), which quantifies weekly energy expenditure in metabolic equivalent task (MET) hours. This tool has shown high test–retest reliability (97%) and acceptable validity (49%) among Iranian children and adolescents [23]. Blood pressure was measured manually with a mercury sphygmomanometer and an appropriately sized cuff after the participant had rested in a seated position for 15 min. Systolic and diastolic blood pressure (SBP and DBP) were determined using the Korotkoff method, with two readings taken at 1‐min intervals and the mean value used for analysis. Biochemical Assessments Venous blood samples were collected between 7:00 and 9:00 a.m. after an overnight fast of 12–14 h. Within 30–45 min of collection, samples were centrifuged at 3000 rpm for 10 min, and all biochemical analyses were performed on the same day. Fasting blood glucose (FBG) and triglyceride (TG) concentrations were measured using standard enzymatic colorimetric assays. Total cholesterol (TC) and high‐density lipoprotein cholesterol (HDL‐C) were determined using cholesterol esterase and phosphotungstic acid precipitation methods, respectively. Low‐density lipoprotein cholesterol (LDL‐C) was estimated using the Friedewald formula [24]. Liver enzyme levels, including aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma‐glutamyl transferase (GGT), were measured using enzymatic techniques. AST and ALT were quantified via enzymatic photometry, while GGT was assessed using a colorimetric assay. All assays were performed using commercial diagnostic kits provided by Delta Darman Inc. (Tehran, Iran) and processed with a Selectra 2 automated analyzer (Vital Scientific, Spankeren, The Netherlands). Fasting serum insulin was measured using an electrochemiluminescence immunoassay (ECLIA) with kits from Roche Diagnostics, analyzed on a Roche/Hitachi Cobas e‐411 system (Roche Diagnostics GmbH, Mannheim, Germany). Insulin resistance (IR) was estimated using the Homeostatic Model Assessment for Insulin Resistance (HOMA‐IR), calculated as (FBS [mg/dL] * Fasting serum insulin [μU/mL])/405. The intra‐ and inter‐assay coefficients of variation for all biochemical parameters were maintained at ≤ 5.3%. Dietary Assessment Dietary intake was evaluated using a validated 147‐item semi‐quantitative FFQ, which has demonstrated both reliability and reproducibility in Iranian populations [25, 26]. Structured, face‐to‐face interviews were conducted by trained nutritionists with each participant and their parent or guardian to assess the frequency and portion sizes of foods consumed over the previous 12 months. In cases where children were unable to accurately report their intake, the mother served as the primary respondent. Portion sizes were primarily based on the United States Department of Agriculture (USDA) food serving standards—for example, one slice of bread, one medium‐sized fruit (e.g., apple), or one cup of milk or yogurt. For local or unlisted foods, standard Iranian household measures were employed (e.g., one tablespoon of legumes, one piece of chicken—leg, breast, or wing—or rice servings categorized as small, medium, or large), and these were subsequently converted to gram weights and standardized servings. Nutrient intake was calculated using the USDA Food Composition Tables (FCTs), which served as the primary reference due to the limited scope of the Iranian FCT, particularly regarding raw foods and beverages. However, for traditional Iranian food items not found in the USDA database—such as Kashk—nutritional values were sourced from the Iranian FCT. The DI‐GM was calculated in accordance with the methodology proposed by Kase et al. (2024), with modifications based on the availability of dietary data in our population [10]. In its original form, the DI‐GM comprises 14 components—10 considered beneficial to gut microbial diversity and 4 considered unfavorable. In our dataset, reliable intake information was available for 12 components: eight beneficial (fermented dairy products, chickpeas, soybeans, whole grains, dietary fiber, cranberries, broccoli, and coffee) and four unfavorable (refined grains, red meat, processed meat, and diets in which total fat accounts for ≥ 40% of total energy intake). Green tea and avocado, present in the original index, were excluded due to the absence of valid consumption data in our dataset. For all components except total fat, dietary intake was expressed in g/day or mL/day, derived from the validated 147‐item FFQ. Total fat intake was calculated as a percentage of total energy. Sex‐specific median intakes for each component, determined from our study population, were used as reference cut‐offs (Supplementary Table 1). For beneficial components, a score of 1 was assigned when intake was at or above the median; otherwise, a score of 0 was recorded. For unfavorable components, scoring was reversed: a value of 1 was assigned when intake was below the median (or < 40% of total energy from fat for the fat component) and 0 otherwise. The DI‐GM score was computed as the sum of these individual scores, ranging from 0 to 12, with higher values reflecting dietary patterns more favorable to gut microbiota diversity. Assessment of MAFLD Hepatic steatosis was assessed after an 8–12‑h fasting period using high‑resolution B‑mode ultrasonography performed by an experienced radiologist who was blinded to participants' clinical data. Imaging was conducted using a Samsung Medison SonoAce R3 system equipped with a 7.5–10 MHz linear transducer. Hepatic steatosis was defined according to standard ultrasonographic criteria, including increased liver echogenicity relative to the renal cortex, attenuation of the ultrasound beam, and reduced visualization of intrahepatic vessels and diaphragm. The diagnosis of MAFLD was established according to the international expert consensus definition. Participants were classified as having MAFLD when hepatic steatosis was present together with overweight or obesity, defined as a BMI‑for‑age Z‑score ≥ 1 based on WHO growth reference standards [3, 19]. Statistical Analysis Descriptive statistics were used to summarize demographic, anthropometric, biochemical, and dietary characteristics of the study population, stratified by MAFLD status and quartiles of the DI‐GM. Continuous variables were assessed for normality using the Kolmogorov–Smirnov test and presented as mean ± SD or median (interquartile range), as appropriate. Between‐group comparisons were performed using independent sample t‐tests or Mann–Whitney U tests for continuous variables and χ 2 tests for categorical variables. Linear regression and χ 2 trend analyses were used to examine trends across DI‐GM quartiles. Associations between DI‐GM and the odds of MAFLD were analyzed using multivariable logistic regression models. Odds ratios (ORs) with 95% confidence intervals (CIs) were reported for each quartile of DI‐GM, using the lowest quartile as the reference category. Covariates were selected based on biological plausibility, empirical evidence from prior epidemiological studies, and their established roles as potential confounders in the association between DI‐GM and metabolic outcomes [11, 12, 13]. The first model adjusted for age and sex. The second model included additional adjustments for pubertal status, BMI‐for‐age Z‐score, PA, TG, and HOMA‐IR. The final model further adjusted for total energy intake. To examine linear trends, the median value of each DI‐GM quartile was included in the model as a continuous variable. Furthermore, the nonlinear dose–response relationship between DI‐GM and the odds of MAFLD was modeled using random‐effects restricted cubic splines, with three knots placed at the 10th, 50th, and 90th percentiles of the DI‐GM distribution [27]. To evaluate the robustness of our findings to the exclusion of avocado and green tea from the DI‐GM, a post hoc sensitivity analysis was conducted using a hypothetical imputation approach. Specifically, random binary variables were generated to simulate plausible consumption of avocado and green tea, based on assumed population‐level probabilities (30% and 50%, respectively). Each participant was assigned a score of 0 or 1 for each component using a Bernoulli distribution, and these values were added to their original DI‐GM score to create an expanded index ranging from 0 to 14. Multivariable logistic regression models were then re‐estimated using DI‐GM as the exposure variable, with the same covariate adjustments applied as in the primary models. Additionally, DI‐GM was categorized into quartiles to examine the consistency of associations across intake levels. This sensitivity analysis allowed for the assessment of whether the original findings were materially influenced by the omission of these two components, thereby strengthening the construct validity of the index within the pediatric population. Mediation analysis was performed to assess whether selected clinical and metabolic biomarkers mediated the association between DI‐GM and MAFLD. The “mediation” package in R was employed to estimate parallel mediation models, evaluating the potential mediating roles of hepatic enzymes (ALT, AST, and GGT), triglycerides, HOMA‐IR, and BMI‐for‐age Z‐score. The proportion mediated was calculated by comparing the indirect effect to the total effect. Confidence intervals were derived using a nonparametric bootstrap procedure with 10,000 resamples, providing robust estimates of mediation effects. All models were adjusted for age, sex, pubertal status, physical activity, and energy intake. Subgroup analyses were additionally performed based on sex (boys vs. girls) and pubertal status (prepubertal vs. pubertal) to explore potential effect modification. A two‐sided p value < 0.05 was considered statistically significant. All statistical analyses were conducted using Stata version 17 (StataCorp, College Station, Texas, the United States) and R software version 4.5.0 (R Foundation for Statistical Computing, Vienna, Austria). Results Study Population Figure 1 presents a detailed flowchart outlining the study population selection process. Of the 548 overweight and obese children and adolescents initially recruited, 31 individuals were excluded due to missing data in key domains, including anthropometric measurements, dietary intake, biochemical parameters, or ultrasonographic assessment. An additional 12 participants were excluded on the basis of implausible dietary reporting. After these exclusions, the final analytical sample consisted of 505 participants. Flowchart of the study. Participant Characteristics Table 1 presents the characteristics of the study participants according to MAFLD status. A total of 505 overweight and obese children and adolescents (mean age: 10.0 ± 2.3 years; 52.9% boys) were included in the analysis. Among them, 196 participants (38.8%) were diagnosed with MAFLD. Compared to their non‐MAFLD counterparts, individuals with MAFLD had significantly higher BMI‐for‐age Z‐scores, WC, body weight, and liver enzyme levels (ALT, AST, and GGT), as well as elevated HOMA‐IR and fasting insulin concentrations (all p < 0.05). Additionally, the DI‐GM score was significantly higher in participants with MAFLD (6.4 ± 1.6) compared to those without MAFLD (7.0 ± 1.7, p = 0.01). Table 1Characteristics of the study participants according to metabolic dysfunction‐associated fatty liver disease status.Total sample (N = 505)Without MAFLD (N = 309)MAFLD (N = 196) p valueDemographic dataAge (years)10.0 ± 2.39.7 ± 2.110.5 ± 2.5 < 0.01 Gender (Boys, %)52.952.154.10.66Puberty (Prepubertal, %)23.423.223.50.96Weight (kg)49.6 ± 15.046.2 ± 12.355.1 ± 17.2 < 0.01 Height (cm)142.2 ± 12.3140.0 ± 11.4145.5 ± 12.9 < 0.01 Body mass index (kg/m2)24.0 ± 3.823.1 ± 3.124.4 ± 4.4 < 0.01 BMI‐for‐age Z‐score2.87 ± 0.982.72 ± 0.802.99 ± 0.72 0.02 Waist circumference (cm)83.5 ± 10.680.9 ± 9.487.6 ± 10.9 < 0.01 Passive smoker (%)23.830.423.90.13Physical activity (MET/h/week)8.6 (3.0–20.4)8.9 (2.6–20.4)7.5 (3.7–20.4)0.89Systolic blood pressure (mmHg)105.0 (97.5–116.0)105.0 (95.0–115.0)105.0 (100.0–120.0)0.26Diastolic blood pressure (mmHg)65.0 (60.0–75.0)65.0 (60.0–75.0)65.0 (60.0–70.0)0.84Biochemical dataFasting serum insulin (mU/mL)16.4 ± 8.615.5 ± 7.917.8 ± 9.6 0.01 Fasting blood sugar (mg/dL)91.1 ± 8.790.6 ± 8.991.9 ± 8.50.11HOMA‐IR3.73 ± 2.163.5 ± 1.94.1 ± 2.3 0.01 Triglyceride (mg/dL)109.5 (81.0–151.5)110.0 (83.0–152.0)168.0 (91.0–122.0) 0.01 Cholesterol (mg/dL)171.0 ± 55.9172.3 ± 66.8169.2 ± 31.80.54HDL (mg/dL)47.1 ± 11.749.0 ± 11.644.0 ± 11.3 < 0.01 LDL‐C (mg/dL)98.0 ± 25.597.3 ± 24.499.1 ± 27.20.43Alanine aminotransferase (U/L)16.0 (11.0–22.0)15.0 (11.0–19.0)18.0 (13.0–32.0) 0.01 Aspartate aminotransferase (U/L)23.0 (17.0–29.0)22.0 (16.0–28.0)25.0 (17.0–32.0) 0.01 Gamma‐glutamyl transferase (U/L)17.0 (15.0–21.0)16.9 (14.0–19.0)19.0 (16.0–24.0) 0.01 Dietary intakeEnergy (kcal/day)3046.3 ± 956.33069.9 ± 998.53009.1 ± 887.00.48Carbohydrate (% of energy)56.0 ± 6.256.1 ± 5.856.0 ± 6.70.89Fat (% of energy)30.6 ± 5.830.6 ± 5.630.4 ± 6.30.76Protein (% of energy)13.4 ± 2.213.3 ± 2.113.6 ± 2.30.20DI‐GM6.8 ± 1.77.0 ± 1.76.4 ± 1.6 0.01 Note: Significant p values are highlighted in bold.Abbreviations: DI‐GM, Dietary Index for Gut Microbiota; HDL, high‐density lipoprotein; HOMA‐IR, Homeostatic Model Assessment for Insulin Resistance; LDL, Low‐density lipoprotein. Characteristics of the study participants according to metabolic dysfunction‐associated fatty liver disease status. Note: Significant p values are highlighted in bold. Abbreviations: DI‐GM, Dietary Index for Gut Microbiota; HDL, high‐density lipoprotein; HOMA‐IR, Homeostatic Model Assessment for Insulin Resistance; LDL, Low‐density lipoprotein. Table 2 represents the clinical, anthropometric, and biochemical characteristics of participants across DI‐GM quartiles. As shown in Table 2, no significant differences were observed in age, sex distribution, pubertal status, or passive smoking exposure across quartiles of the DI‐GM. Anthropometric measures, including BMI‐for‐age Z‐score, WC, and body weight, also did not differ significantly across quartiles. Serum levels of ALT, AST, and GGT showed statistically significant differences across DI‐GM quartiles (p < 0.05 for all). No significant trends were noted for FBS, insulin, HOMA‐IR, TC, HDL‐C, LDL‐C, TG, or blood pressure measurements. Physical activity levels (METs/week) and the proportion of participants with MAFLD were also not significantly different across quartiles. Table 2Characteristics of the participants across quartiles of the Dietary Index for Gut Microbiota* , ‡ .Q1Q2Q3Q4 p valueAge (years)10.1 ± 2.49.9 ± 2.2.10.1 ± 2.209.5 ± 2.10.19Sex (male, %)111 (55.0)65 (53.8)49 (48.0)43 (52.4)0.72Pubertal status (prepubertal %)52 (25.7)27 (22.7)22 (21.6)17 (20.7)0.76Height (cm)142.3 ± 13.0142.3 ± 12.0143.4 ± 12.0140.1 ± 11.00.42Weight (kg)50.0 ± 15.750.2 ± 15.550.6 ± 14.446.7 ± 13.10.21Waist circumference (cm)83.9 ± 11.183.7 ± 10.983.9 ± 9.281.8 ± 10.30.23Body mass index (kg/m2)24.0 ± 3.824.4 ± 4.424.2 ± 3.423.3 ± 3.40.26Body mass index for age Z‐score2.85 ± 1.02.94 ± 1.02.86 ± 0.892.78 ± 0.860.63Physical activity (MET/h/week)7.9 (3.1–20.5)6.4 (1.9–18.8)9.1 (2.2–17.5)12.8 (4.8–22.7)0.20Passive smoker (%)58 (28.7)31 (26.1)21 (20.6)24 (29.3)0.44Fasting blood sugar (mg/dL)91.6 ± 8.791.1 ± 8.691.3 ± 9.489.9 ± 8.40.20Fasting serum insulin (μU/mL)16.8 ± 9.316.5 ± 8.615.9 ± 7.415.7 ± 8.50.31Homeostatic Model Assessment for Insulin Resistance3.87 ± 2.383.78 ± 2.183.58 ± 1.753.48 ± 1.970.16Total cholesterol (mg/dL)170.3 ± 31.5165.9 ± 31.9166.7 ± 29.2185.7 ± 29.00.13High‐density lipoprotein cholesterol (mg/dL)47.0 ± 11.747.2 ± 11.847.4 ± 12.046.9 ± 11.70.97Low‐density lipoprotein cholesterol (mg/dL)100.5 ± 25.393.6 ± 25.097.8 ± 24.498.5 ± 27.90.48Triglycerides (mg/dL)110.5 (88.5–152.0)112.0 (82.0–153.0)117.0 (84.0–165.5)114.0 (88.0–161.0)0.31Alanine aminotransferase (IU/L)16.0 (12.0–24.0)17.0 (12.0–24.0)14.0 (11.0–19.0)17.0 (10.0–21.2) < 0.01 Aspartate aminotransferase (IU/L)25.0 (17.0–30.2)24.0 (16.0–30.0)21.0 (16.0–27.0)24.0 (25.7–29.0) 0.04 Gamma‐glutamyl transferase (IU/L)18.0 (15.0–21.0)18.0 (16.0–22.0)17.0 (14.5–19.0)17.0 (13.9–20.0) 0.01 Systolic blood pressure (mmHg)105 (96–115)105 (98–117)105 (100–120)107 (95–117)0.71Diastolic blood pressure (mmHg)68 (60–71)61 (60–70)65 (60–70)65 (60–80)0.28*All values are reported as mean ± standard deviation or median (quartiles 25–75) and qualitative variables as percentages. ‡ Linear regression and χ 2 analyses were used to test the trend of continuous and categorical variables, respectively, according to quartiles of the Dietary Index for Gut Microbiota. Significant p values are highlighted in bold. Characteristics of the participants across quartiles of the Dietary Index for Gut Microbiota* , ‡ . All values are reported as mean ± standard deviation or median (quartiles 25–75) and qualitative variables as percentages. Linear regression and χ 2 analyses were used to test the trend of continuous and categorical variables, respectively, according to quartiles of the Dietary Index for Gut Microbiota. Significant p values are highlighted in bold. Table 3 summarizes the distribution of dietary intake variables across DI‐GM quartiles. Total energy intake was significantly higher among individuals in the top quartiles of DI‐GM (p value < 0.01). The percentage of energy derived from carbohydrates showed a significant upward trend, while the proportions of energy from total fat, saturated fat (SFA), and monounsaturated fat (MUFA) (p < 0.001) declined significantly across increasing DI‐GM quartiles (all p values < 0.01). No significant differences were noted in protein intake across groups. Food group analysis revealed significant trends across DI‐GM quartiles. Higher quartiles were associated with greater consumption of whole grains, legumes, fruits, and vegetables and lower intake of refined grains and red meat (all p values < 0.01). Micronutrient and fiber intake also varied significantly. Participants in the higher DI‐GM quartiles had greater intakes of dietary fiber, calcium, iron, zinc, and magnesium (all p values < 0.01). Table 3Dietary intakes of the participants across quartiles of the Dietary Index for Gut Microbiota* , ‡ .Q1Q2Q3Q4 p valueEnergy (kcal)2904 ± 9243021 ± 8983137 ± 10513319 ± 937 < 0.01 Carbohydrate (% of energy)55.0 ± 6.256.2 ± 7.057.2 ± 5.857.1 ± 4.8 < 0.01 Protein (% of energy)13.4 ± 2.413.4 ± 2.313.4 ± 1.813.7 ± 1.80.31Fat (% of energy)34.2 ± 5.832.7 ± 6.532.2 ± 5.731.9 ± 4.4 < 0.01 Whole grain (serving/day)1.6 (0.80–4.2)1.7 (1.0–3.5)2.7 (1.3–5.8)3.3 (1.6–7.0) < 0.01 Refined grain (serving/day)11.5 (8.7–15.4)11.0 (7.6–15.4)10.9 (7.9–14.4)10.3 (7.7–12.6) 0.02 Red meat (serving/day)0.71 (0.32–1.15)0.57 (0.26–1.14)0.43 (0.20–0.84)0.44 (0.15–1.06) 0.02 White meat (serving/day)1.16 (0.65–2.05)1.33 (0.65–2.36)1.25 (0.66–1.66)1.33 (0.85–1.83)0.11Dairy (serving/day)2.0 (1.3–3.1)2.3 (1.3–3.3)2.4 (1.5–3.2)2.5 (1.5–3.5)0.27Legumes (serving/day)0.30 (0.17–0.55)0.44 (0.26–0.72)0.54 (0.34–0.87)0.65 (0.44–1.00) < 0.001 Nuts (serving/day)0.31 (0.12–0.58)0.34 (0.14–0.75)0.33 (0.17–0.87)0.33 (0.21–0.61)0.14Fruits (serving/day)3.4 (2.3–4.9)3.8 (2.7–5.8)4.3 (2.9–6.6)4.5 (3.3–6.2) < 0.001 Vegetables (serving/day)2.2 (1.4–3.0)2.5 (1.5–4.0)2.7 (2.1–4.4)3.8 (2.7–5.4) < 0.001 SFA (% of energy)10.9 ± 2.710.3 ± 2.810.3 ± 2.310.2 ± 2.2 0.04 MUFA (% of energy)11.1 ± 2.510.2 ± 2.410.5 ± 2.410.2 ± 1.7 < 0.01 PUFA (% of energy)7.0 ± 2.06.9 ± 2.16.8 ± 1.86.5 ± 1.50 0.04 Sodium (mg/2000 kcal)2853 ± 7452840 ± 9532890 ± 6892917 ± 7310.49Calcium (mg/2000 kcal)918 ± 283940 ± 266965 ± 2551005 ± 271 0.01 Zinc (mg/2000 kcal)9.1 ± 1.79.3 ± 1.59.6 ± 1.89.9 ± 1.6 < 0.01 Magnesium (mg/2000 kcal)294 ± 54308 ± 48331 ± 60348 ± 59 < 0.001 Iron (mg/2000 kcal)13.2 ± 2.213.5 ± 2.713.9 ± 2.314.8 ± 3.1 < 0.001 Fiber (g/2000 kcal)31.8 ± 11.935.2 ± 13.735.3 ± 10.638.2 ± 10.1 < 0.001 Abbreviations: MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids.*All values are reported as mean ± standard deviation. ‡ Linear regression analyses were used to test the trend of continuous variables according to quartiles of the Dietary Index for Gut Microbiota. Significant p values are highlighted in bold. Dietary intakes of the participants across quartiles of the Dietary Index for Gut Microbiota* , ‡ . Abbreviations: MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids. All values are reported as mean ± standard deviation. Linear regression analyses were used to test the trend of continuous variables according to quartiles of the Dietary Index for Gut Microbiota. Significant p values are highlighted in bold. Table 4 summarizes the ORs and 95% CIs for the association between DI‐GM and MAFLD across quartiles, per unit, and per SD increase. In the unadjusted model, participants in the third quartile had an OR of 0.37 (95% CI: 0.20–0.67), and those in the fourth quartile had an OR of 0.42 (95% CI: 0.23–0.76), compared to those in the lowest quartile. After adjusting for age and sex (Model 1), the ORs for Q3 and Q4 were 0.38 (95% CI: 0.20–0.70) and 0.44 (95% CI: 0.24–0.82), respectively. In Model 2, which included additional adjustments for pubertal status, BMI‐for‐age Z‐score, physical activity, triglycerides, and HOMA‐IR, the inverse associations persisted with ORs of 0.35 (95% CI: 0.18–0.65) for Q3 and 0.42 (95% CI: 0.22–0.81) for Q4. These associations remained significant in the fully adjusted model (Model 3), which further controlled for total energy intake. Trend analysis across quartiles revealed a significant linear inverse association in all models (p for trend < 0.001). Table 4Odds ratios (95% confidence intervals) for metabolic dysfunction‐associated fatty liver disease according to the proportion of Dietary Index for Gut Microbiota score.Q1Q2Q3Q4 p trendContinues per one unit increment p valueContinues per one SD increment p valueCases/populations94/20253/11926/10223/82Median score5789Crude model1 (Ref)0.92 (0.58–1.45) 0.39 (0.23–0.66) 0.45 (0.26–0.78) < 0.001 0.80 (0.73–0.89) < 0.001 0.69 (0.57–0.82) < 0.001 Model 1a 1 (Ref)0.94 (0.59–1.49) 0.38 (0.22–0.65) 0.47 (0.27–0.83) < 0.001 0.80 (0.72–0.89) < 0.001 0.69 (0.57–0.83) < 0.001 Model 2b 1 (Ref)0.77 (0.46–1.30) 0.34 (0.18–0.65) 0.41 (0.22–0.79) < 0.01 0.80 (0.71–0.90) < 0.001 0.68 (0.56–0.84) < 0.001 Model 3c 1 (Ref)0.77 (0.46–1.30) 0.35 (0.18–0.65) 0.42 (0.22–0.81) < 0.001 0.80 (0.71–0.91) < 0.001 0.69 (0.56–0.85) < 0.001 Note: Significant p values are highlighted in bold. a Model 1 adjusted for age and sex. b Model 2 additionally adjusted for Model 1 and puberty status, body mass index for age Z‐score, physical activity, triglycerides, and the Homeostatic Model Assessment for Insulin Resistance. c Model 3 additionally adjusted for Model 2 and dietary energy intake. Odds ratios (95% confidence intervals) for metabolic dysfunction‐associated fatty liver disease according to the proportion of Dietary Index for Gut Microbiota score. Note: Significant p values are highlighted in bold. Model 1 adjusted for age and sex. Model 2 additionally adjusted for Model 1 and puberty status, body mass index for age Z‐score, physical activity, triglycerides, and the Homeostatic Model Assessment for Insulin Resistance. Model 3 additionally adjusted for Model 2 and dietary energy intake. In the continuous analysis, each one‐unit increase in DI‐GM was associated with a 20% reduction in the odds of MAFLD (OR = 0.80, 95% CI: 0.71–0.91; p value < 0.001). Moreover, when modeled per SD increase in DI‐GM, the association remained consistent, with a 31% reduction in MAFLD odds observed in the fully adjusted model (OR = 0.69, 95% CI: 0.56–0.85; p < 0.001). Supplementary Table 2 presents ORs and 95% CIs for MAFLD according to the imputed DI‐GM score, which incorporates simulated intake of avocado and green tea. The inverse association between imputed DI‐GM and MAFLD remained robust across all models. In the fully adjusted model, participants in the highest quartile had a 53% lower odds of MAFLD compared to those in the lowest quartile (OR = 0.47, 95% CI: 0.24–0.93; p trend = 0.01). In continuous analyses, each one‐unit increase in imputed DI‐GM was significantly associated with a 14% reduction in the odds of MAFLD (OR = 0.86, 95% CI: 0.77–0.96; p value < 0.01). Similarly, each SD increase in imputed DI‐GM was significantly associated with a 24% lower odds of MAFLD (OR = 0.76, 95% CI: 0.62–0.93; p < 0.01). These findings confirm the robustness of the original results and demonstrate that the exclusion of avocado and green tea in the initial index did not materially alter the observed associations. Table 5 presents the results of the subgroup analysis assessing the association between the DI‐GM and the odds of MAFLD, stratified by gender and pubertal status. In both boys and girls, higher DI‐GM scores were associated with significantly lower odds of MAFLD, particularly among those in the third and fourth quartiles. Among boys, the odds of MAFLD were significantly reduced in Q3 (OR = 0.38, 95% CI: 0.15–0.98) and Q4 (OR = 0.38, 95% CI: 0.15–0.94; p trend = 0.01), with a statistically significant interaction between DI‐GM and gender (p value for interaction < 0.01). Similarly, girls in Q3 had substantially lower odds of MAFLD (OR = 0.29, 95% CI: 0.12–0.72), although the association in Q4 did not reach statistical significance (OR = 0.46, 95% CI: 0.18–1.17). When stratified by pubertal status, a clear inverse association was observed only among pubertal participants. In this group, higher DI‐GM quartiles were significantly associated with lower odds of MAFLD (Q3: OR = 0.36, 95% CI: 0.18–0.71; Q4: OR = 0.40, 95% CI: 0.19–0.82; p for trend = 0.01), while no significant association was detected in the prepubertal subgroup (p trend = 0.20). A significant interaction was also observed between DI‐GM and pubertal status (p value for interaction < 0.01). Table 5Subgroup analysis of the association between the Dietary Index for Gut Microbiota score and the odds of metabolic dysfunction‐associated fatty liver disease a .Quartile of the Dietary Index for Gut Microbiota scoreQ1Q2Q3Q4 p trend p value for interactionGenderBoys (N = 267)1.00 (Ref)0.92 (0.45–1.90) 0.38 (0.15–0.98) 0.38 (0.15–0.94) 0.01 < 0.01 Girls (N = 238)1.00 (Ref)0.56 (0.26–1.24) 0.29 (0.12–0.72) 0.46 (0.18–1.17) 0.01 Puberty statusPrepubertal (N = 118)1.00 (Ref)0.41 (0.09–1.73)0.12 (0.01–1.34)0.75 (0.12–4.69)0.20 < 0.01 Pubertal (N = 387)1.00 (Ref)0.73 (0.40–1.33) 0.36 (0.18–0.71) 0.40 (0.19–0.82) 0.01 Note: Significant p values are highlighted in bold.Abbreviations: HDL, high‐density lipoprotein; HOMA‐IR, Homeostatic Model Assessment for Insulin Resistance. a Odds ratios (95% CI) for metabolic dysfunction‐associated fatty liver disease were obtained from multivariate logistic regression analysis, adjusting for age, sex, puberty status, body mass index for age Z‐score, physical activity, triglycerides, Homeostatic Model Assessment for Insulin Resistance, systolic blood pressure, and dietary energy intake. Subgroup analysis of the association between the Dietary Index for Gut Microbiota score and the odds of metabolic dysfunction‐associated fatty liver disease a . Note: Significant p values are highlighted in bold. Abbreviations: HDL, high‐density lipoprotein; HOMA‐IR, Homeostatic Model Assessment for Insulin Resistance. Odds ratios (95% CI) for metabolic dysfunction‐associated fatty liver disease were obtained from multivariate logistic regression analysis, adjusting for age, sex, puberty status, body mass index for age Z‐score, physical activity, triglycerides, Homeostatic Model Assessment for Insulin Resistance, systolic blood pressure, and dietary energy intake. Figure 2 displays the restricted cubic spline curve representing the nonlinear association between DI‐GM and the odds of MAFLD. The model, adjusted for age, sex, pubertal status, BMI‐for‐age Z‐score, physical activity, triglycerides, HOMA‐IR, and total energy intake, showed a statistically significant inverse association (P for nonlinearity = 0.03). The steepest decline in MAFLD odds was observed within the lower to mid‐range of the DI‐GM score distribution. Multivariable adjusted spline curve for the association between the Dietary Index for Gut Microbiota and the odds of metabolic dysfunction‐associated fatty liver disease (MAFLD). Figure 3 presents the results of the parallel mediation analysis exploring the potential indirect associations of selected biomarkers with the relationship between DI‐GM and MAFLD. The analysis indicated that the association between DI‐GM and MAFLD was significantly mediated by ALT, AST, and GGT. Specifically, ALT mediated 20.0% (p < 0.001), GGT mediated 19.2% (p = 0.001), and AST mediated 6.7% (p = 0.03) of the total estimated indirect association. HOMA‐IR showed a borderline significant mediation effect (10.7%, p = 0.06), while triglycerides and BMI‐for‐age Z‐score did not significantly mediate the relationship. Mediating role of clinical and metabolic intermediates in the association between the Dietary Index for Gut Microbiota and the odds of metabolic dysfunction‐associated fatty liver disease (MAFLD). (A) Alanine aminotransferase; (B) aspartate aminotransferase; (C) gamma‐glutamyl transferase; (D) triglyceride; (E) Homeostatic Model Assessment for Insulin Resistance (HOMA‐IR); and (F) body mass index for age Z‐score. Fully adjusted for age, sex, puberty status, physical activity, and dietary energy intake. Discussion This is the first study to evaluate the association between DI‐GM and the odds of MALFD among overweight and obese children and adolescents. Our findings showed that a higher DI‐GM was linked to lower odds of developing MAFLD in this population, after adjusting for potential confounders. The fully adjusted model included age, sex, pubertal status, BMI‑for‑age Z‑score, physical activity, triglycerides, HOMA‑IR, and total energy intake. This association was evident in both continuous (OR: 0.80, 95% Cl: 0.71–0.91, p < 0.001) and highest versus lowest models (Quartile 4 vs. Quartile 1, OR: 0.42, 95% Cl: 0.22–0.81, p trend< 0.001) within the fully adjusted model. Mediation analysis identified ALT (proportion of mediation = 20.0%, p value < 0.001), GGT (proportion of mediation = 19.2%, p value < 0.01), and AST (proportion of mediation = 6.7%, p value = 0.03) as significant mediators in the relationship between gut microbiota‐related dietary scores and the odds of MAFLD. The gut microbiome should be regarded as an organ system that significantly influences childhood development. Research indicates that the development of infants' gut microbiota is highly predictable, closely linked to their exposures (such as diet), and serves as a predictor of future health outcomes. This relationship is so crucial that statistical modeling has been employed to quantify how early‐life microbiome characteristics affect an individual's long‐term health [28]. The initial hypothesis connecting obesity to gut microbiota emerged from studies involving germ‐free (GF) mice. Notably, mice raised in a typical environment exhibited a 40% higher fat content compared to those reared in a sterile setting. Remarkably, the transplantation of gut microbiota from conventional mice into GF mice led to a 60% increase in body fat within just 14 days, without any increase in food intake or significant changes in energy expenditure [29]. Various theories suggest that the gut microbiome in children plays a significant role in the development of obesity and metabolic disorders. Typically, a state of normobiosis includes beneficial strains like Bacteroides, Bifidobacterium, Lactobacillus, Prevotella, Ruminococcus, Christensenella, Roseburia, and Akkermansia. This diverse microbiome promotes the production of short‐chain fatty acids (SCFAs), enhances insulin sensitivity, and reduces inflammation. In contrast, dysbiosis is characterized by strains such as Firmicutes, Bacteroides, Acinetobacter, Proteobacteria, Sutterella, and Propionibacterium, which can lead to increased inflammation, metabolic endotoxemia, and intestinal permeability, along with diminished gut diversity [30]. More straightforwardly, the Firmicutes/Bacteroidetes ratio has frequently been regarded as a potential predictor of obesity risk, particularly in childhood [31]. Research on obesity prevalence and microbiome diversity linked to various dietary patterns has yielded significant insights. Adopting a predominantly plant‐based diet has been associated with an increase in the abundance of Prevotella, Bacteroides, and Treponema, resulting in an obesity prevalence of less than 5% [32]. In contrast, the Western diet is characterized by a decrease in Prevotella and an increase in Bacteroides, with obesity prevalence rates of 38.2% in the US population, 16.7% in Spain, 12.8% in the Netherlands, and 9.8% in Italy [33, 34], all of which surpass those of plant‐based diet consumers. A diet rich in plant‐based foods has been shown to enhance the relative abundance of butyrate‐producing bacteria, such as Faecalibacterium prausnitzii and Roseburia, which are recognized for their protective roles against cardiometabolic diseases [35]. Overall, a review of existing evidence indicates that the regulation of microbiome diversity through plant‐based proteins increases SCFAs and reduces inflammation within the body, whereas animal proteins tend to decrease SCFA levels. Furthermore, unsaturated fats have been shown to negatively impact toll‐like receptors (TLRs) and white adipose tissue (WAT) due to microbiome alterations, while saturated fats are linked to decreased insulin sensitivity. Unsaturated fats tend to support a healthy gut microbiota, while high saturated‑fat intake is linked to dysbiosis, inflammation, and impaired insulin sensitivity [36]. Enterodiol and enterolactone are phytoestrogens, specifically classified as mammalian lignans. They are produced in the colon by gut bacteria from plant‐derived precursors such as secoisolariciresinol and matairesinol [37]. In the research conducted by Kase et al., urinary enterolignans and enterolactone were chosen as outcomes, and their levels demonstrated how diet influences the gut microbiome [10]. Notably, the DI‐GM showed a stronger association with these urinary markers compared to other well‐established dietary patterns, including the Mediterranean diet (7). The gut–liver axis is critically important, as evidence has shown that this dietary pattern lowers the risk of NAFLD [11], MAFLD [12], and MetS [13]. Additionally, hyperlipidemia risk was reduced by 40% in individuals with a DI‐GM greater than 6 compared to those with a DI‐GM less than 3 [38]. Another study indicated that when the DI‐GM fell below 6.191, the risk of diabetes decreased by 7.9% [39]. Furthermore, each incremental increase in the DI‐GM score corresponded with a 7% reduction in diabetes risk. Interestingly, phenotypic age—an aging indicator based on biomarkers reflecting an individual's health status and disease risk—and BMI accounted for 41.02% and 25.57% of the association between DI‐GM and diabetes, respectively [40]. These metabolic changes linked to DI‐GM may serve as promising indicators for reducing the risk of cardiovascular disease (CVD) [41, 42] and chronic kidney disease (CKD) [43]. Previous studies have identified several cut‐off points for DI‐GM in relation to adverse metabolic outcomes. A DI‐GM score of 5 or above reduces the risk of NAFLD [11] and MetS [13]. Scores of 6 or higher are linked to a lower risk of MAFLD [12] and hyperlipidemia [38]. In contrast, the findings of the present study suggest that a higher threshold, a DI‐GM score of 8, may be necessary to identify MAFLD odds among obese children and adolescents. This may imply that greater dietary adherence is required in pediatric populations to achieve similar metabolic benefits. Given the developmental and physiological differences in this age group, more studies are warranted to validate these thresholds and to explore age‐specific dietary strategies for MAFLD prevention and management in youth. Although higher DI‐GM scores were associated with lower odds of MAFLD in both the continuous models and the highest‐versus‐lowest quartile comparisons, this inverse association warrants careful interpretation. In continuous models, each unit and each SD increase in DI‐GM was significantly associated with reduced odds of MAFLD. However, an unexpected pattern emerged in the quartile analysis, where Q3 showed lower odds of MAFLD compared to Q4, despite the overall inverse trend across quartiles. This nonlinear pattern may reflect random variation or other unmeasured factors and, therefore, should be interpreted with caution. Moreover, dietary analysis revealed that participants in Q4 had higher total energy intake compared with other quartiles, which could attenuate the potential benefits of a microbiota‐supportive dietary pattern. It is also plausible that residual confounding from unmeasured factors (such as differences in food preparation methods, nutrient bioavailability, or lifestyle characteristics including sleep quality, stress levels, or genetic predisposition) may have contributed to this finding. Given these possibilities, further studies with larger sample sizes, detailed dietary profiling, and comprehensive assessment of lifestyle and metabolic variables are warranted to elucidate the underlying mechanisms and confirm whether this pattern represents a true biological phenomenon or a statistical artifact. Subgroup analysis indicated that the association between DI‐GM and MAFLD remained statistically significant only among individuals who had reached pubertal status. The prevalence of MAFLD is generally higher post‐puberty, potentially due to hormonal changes that characterize this developmental stage [44]. Puberty is accompanied by marked increases in sex steroids (estrogens and androgens), growth hormone (GH), and IGF‑1, which have been shown to influence gut‑microbiota composition by altering bile‑acid metabolism, mucosal immunity, and intestinal barrier function. These hormonal shifts occur alongside elevated serum insulin levels and increased insulin resistance, all of which may contribute to hepatic fat accumulation [45]. Notably, the interaction terms for both pubertal status and gender were statistically significant (p for interaction < 0.05). Gut‐microbiota composition changes throughout childhood and adolescence [46, 47]. Evidence suggests that microbial diversity and structure evolve with age and undergo additional shifts during puberty due to hormonal, dietary, and physiological changes [47]. Compared with pre‑pubertal children, adolescents generally exhibit greater microbial diversity and differences in taxa involved in metabolic regulation [46]. These developmental changes may partly explain why the inverse association between DI‑GM and MAFLD was observed mainly in post‑pubertal participants in the present study. An investigation into dietary items across the DI‐GM quartiles found that higher quartiles had greater energy intake from carbohydrates and less from fats. This suggests that a diet rich in complex and healthy carbohydrates may lower the risk of MAFLD. A related case–control study showed that low‐carb diets were not significantly linked to a reduced risk of NAFLD [48]. Additionally, diets low in saturated fat and high in complex carbohydrates, like the Mediterranean diet, were associated with lower intrahepatic triglycerides and better insulin sensitivity [49]. The current study also indicated that adherence to DI‐GM led to increased consumption of whole grains, legumes, fruits, and vegetables, while decreasing red meat intake, a key source of dietary cholesterol. These dietary patterns reflect a shift toward anti‐inflammatory and fiber‐rich foods that may play a protective role against hepatic fat accumulation and metabolic disturbances in children and adolescents. Reduction of inflammation is the most likely mechanism by which a diet that positively influences the gut microbiome may help decrease the risk of MAFLD. A recent study identified the SII and NLR as mediators of the relationship between DI‐GM and MetS [13]. Both of these inflammatory markers, derived from blood tests, reflect the balance between various types of white blood cells and can indicate the presence and severity of systemic inflammation [50]. However, a more specific understanding of the connection between this dietary pattern and MASLD is needed. The current study demonstrated that ALT, GGT, and AST play significant roles in mediating this association. Under normal conditions, only a small percentage of gut microbiota components enter the liver due to the intestinal barrier. However, in patients with NAFLD, gut dysbiosis disrupts this barrier, allowing more gut‐derived components to reach the liver via the portal vein. This translocation can alter liver function and affect traditional biomarkers like ALT, GGT, and AST, which are elevated in liver damage [51]. An animal study explored the relationship between fast food diets and changes in the gut microbiome in mice. Researchers divided the mice into two groups: conventionally raised mice and those raised in a GF environment, completely devoid of all microbiota. The results showed that ALT levels were significantly higher in mice on a high‐fat diet compared to those on a low‐fat diet, but this was observed only in the conventionally raised mice and not in the GF mice. This indicates that microbial presence has a strong influence on ALT levels, highlighting its important mediating role between diet and the development of NAFLD [52]. The current study also aligns with previous findings that liver enzymes, including ALT, GGT, and AST, mainly mediated the link between diet and MAFLD, indicating that improved hepatic enzyme profiles may be a key mechanism through which diet influences liver health and its related metabolic dysfunction. The DI‐GM dietary pattern is considered healthy and linked to improved liver function. Randomized clinical trials provide evidence that certain dietary factors contribute to these beneficial outcomes. For instance, a 12‐week study found that consuming 98 g of whole‐grain wheat daily prevented an increase in liver fat, while gut microbiota diversity decreased in those eating refined wheat [53]. Another 12‐week trial with NAFLD patients showed that those who consumed half their cereal servings from whole grains had lower ALT, AST, and GGT levels compared to those who ate regular cereals [54]. It is important to note that the current study did not identify HOMA‐I, TG, or BMI‐for‐age Z‐scores as mediators (p < 0.05). However, other studies focusing on adults found that BMI mediated the relationship between DI‐GM and atherosclerotic cardiovascular disease (ASCVD) (11.51%) [41], infertility (5.98%) [55], and diabetes (25.57%) [40]. In children, BMI‐for‐age Z‐scores may be less responsive to short‐term dietary changes compared to adults, which could explain the nonsignificant findings in this study. Additionally, the lack of clear mediators may also be influenced by specific demographic population groups and small sample sizes. These findings underscore the complexity of diet‐health associations in pediatric populations and highlight the need for further large‐scale, age‐specific mediation analyses. Strengths and Limitations The current study demonstrates several strengths. Firstly, its focus on MAFLD within a substantial sample of the obese pediatric population—a particularly sensitive demographic—may aid in the development of strategies aimed at reducing obesity and metabolic diseases in adulthood. Employing validated questionnaires to evaluate dietary intake and physical activity guarantees the strength of the data. The involvement of mothers during in‐person interviews probably enhanced the precision of dietary recall and the quantification of intake among child participants. The use of mediation analysis is another advantage, highlighting the importance of traditional liver function markers that play a significant role in mediating the association between DI‐GM and MAFLD, though its findings need to be confirmed in longitudinal studies. Although DI‐GM is a newly emerged dietary pattern that shows promise for promoting healthier eating habits by considering changes in the gut microbiome, it is important to acknowledge that enterodiol and enterolactone were utilized as indirect biomarkers of gut microbiota diversity during the development process [10]. Future research with datasets providing direct measures of gut microbiota diversity may offer clearer insights. Despite its strengths, this study has several limitations that warrant consideration. First, the cross‐sectional design precludes any inference of causality between dietary patterns, gut microbiota‐related effects, and MAFLD outcomes. Longitudinal or intervention‐based studies are needed to establish temporal relationships and directionality. Second, although the FFQ was validated for the target population, self‐reported dietary data—especially in children—are prone to recall bias and misreporting, even with caregiver involvement; such measurement errors may affect the validity of the findings. Residual confounding may persist due to unmeasured socioeconomic factors and potential inaccuracies in dietary reporting, which could influence both exposure and outcome measures. Third, the absence of data on certain food items known to influence gut microbiota (such as avocado and green tea) due to their low prevalence in the local diet may lead to underestimation of their potential effects on DI‐GM scores. However, even when a post hoc sensitivity analysis was conducted using a hypothetical imputation approach to account for this missing data, the statistical significance of the findings remained unchanged. Fourth, although several potential confounders were accounted for in our analytical models, socioeconomic and environmental factors may still influence both dietary patterns and liver health. Fifth, while the DI‐GM is a promising dietary index grounded in microbiome research, it does not incorporate direct microbiota sequencing data. Instead, its development relied on surrogate biomarkers such as enterodiol and enterolactone, which limit its precision in capturing microbial diversity and function. Sixth, the study sample consisted exclusively of overweight and obese children and adolescents, which, while relevant to the condition under investigation, may restrict the generalizability of findings to normal‐weight or underweight populations. Seventh, although liver biopsy is the gold standard for diagnosing hepatic steatosis, this study used ultrasonography—a widely used, noninvasive method that lacks sensitivity for mild steatosis and cannot evaluate fibrosis. Eighth, although the DI‐GM was originally developed and validated in adult populations, its direct applicability to children has not yet been established through validation against microbiome measures in this age group. While our findings suggest that the DI‐GM may have utility in pediatric settings, differences in dietary patterns, physiological development, and gut‐microbiota composition across life stages may influence its precision in capturing diet–microbiota relationships in children and adolescents. This underscores the need for future research incorporating direct microbiome profiling to validate and, if necessary, recalibrate the index for pediatric populations, thereby ensuring its accuracy and relevance in early life stages. Finally, antibiotic use was not assessed, which may influence the observed association between DI‐GM and the gut microbiome. Conclusion This cross‐sectional study, the first of its kind conducted in a pediatric population, revealed that adherence to a dietary pattern that positively influences the gut microbiome, as indicated by higher scores in the DI‐GM, is associated with a reduced risk of MAFLD in obese children and adolescents. The relationship is primarily mediated by changes in ALT, GGT, and AST. These findings suggest that dietary patterns influencing gut microbiome dynamics may contribute to improved health outcomes, though further research with direct microbiota assessment is needed to confirm this relationship. Author Contributions Overall, Golaleh Asghari and Pejman Rohani supervised the project and approved the final version of the manuscript to be submitted. Ali Nikparast designed the research. Ali Nikparast and Elahe Etesami analyzed and interpreted the data. Farzam Kamrani and Ali Tabatabaeyan drafted the initial manuscript. and Golaleh Asghari critically revised the manuscript. All authors approved the final version of the manuscript submitted for publication. Ethics Statement National Nutrition and Food Technology Research Institute (NNFTRI) ethics committee approved the study protocol (IR.SBMU.NNFTRI.REC.1402.015). All procedures performed in studies involving human participants adhered to the ethical standards of the institutional and/or national research committee and to the 1964 Helsinki declaration and its later amendments or comparable ethical standards. All authors have read and approved the final version of the manuscript. Golaleh Asghari had full access to all data in this study and takes full responsibility for the integrity of the data and the accuracy of the data analysis. The corresponding authors affirm that this manuscript is an honest, accurate, and transparent account of the study being reported and that no important aspects of the study have been omitted. Consent All participants provided written informed consent and were informed about the study. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supporting File: hsr272640‐sup‐0001‐Supplementary_file.docx. Acknowledgments The authors express their appreciation to the participants of the study for their enthusiastic support and to the staff of the involved hospitals for their valuable help. This study is taken from the Obesity Registry Program in children at the Tehran University of Medical Sciences (IR.TUMS.CHMC.REC.1401.016). We are thankful to Dr. Mohammad Hassan Sohouli and Dr. Afshin Ostovar, Head of the Obesity Registry at the Tehran University of Medical Sciences. Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. References Associated Data Supplementary Materials Supporting File: hsr272640‐sup‐0001‐Supplementary_file.docx. Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.