Determinants of Liver Steatosis Progression in Chinese Children: A Prospective Cohort Study

Key Points

• We established a prospective cohort of Chinese pediatric patients to investigate the determinants of longitudinal changes in the controlled attenuation parameter (CAP) and the progression of liver steatosis.
• Biomarkers associated with changes in CAP can be categorized into three groups: two groups are primarily characterized by metabolic and inflammatory biomarkers, respectively, while the neutrophil-to-albumin ratio forms a separate cluster.
• Four independent determinants of steatosis progression (sex, BMI z-score, lymphocyte count, and trunk body fat mass) were identified, with the model achieving good discriminatory ability (AUC = 0.80).

Introduction

Metabolic dysfunction–associated steatotic liver disease (MASLD) has emerged as the leading cause of chronic liver disease in children, with a reported global prevalence ranging from 7% to 35%.^1,2 MASLD significantly increases the risk of cardiovascular events, chronic kidney disease, malignancies both within and beyond the liver, and other various adverse liver-related outcomes, imposing a considerable societal and economic burden.^3,4 Therefore, early diagnosis is of paramount importance.⁵

Currently, the diagnosis of pediatric fatty liver disease primarily relies on imaging techniques, with ultrasound being the preferred initial screening tool due to its cost-effectiveness and accessibility.^6,7 However, conventional ultrasound permits only a qualitative assessment of steatosis, which restricts its diagnostic accuracy.⁵ In contrast, magnetic resonance proton density fat fraction provides a more precise quantification of liver fat content; however, its clinical application is limited by high cost and restricted availability.⁸ In light of these limitations, the controlled attenuation parameter (CAP) has emerged as a noninvasive, reproducible, and quantitative tool that is now widely utilized in pediatric populations.^9,10

The CAP has emerged as a significant indicator for assessing hepatic steatosis. Previous studies have identified several factors that may influence variations in CAP values.¹¹ For instance, a population-based study conducted in the United States (N = 4870) demonstrated that CAP values were a positive association between CAP values and alanine aminotransferase (ALT) levels, body mass index (BMI), and body composition indicators such as fat mass.¹² Another cross-sectional study involving 1080 adolescents found a similar positive correlation between CAP values and BMI, waist-hip ratio (WHR), total fat percentage, and trunk fat percentage.¹³ Likewise, a cohort study in Milan, Italy (Liver‐bible‐2022 cohort, N = 1230) identified abdominal adiposity and the severity of insulin resistance as significant determinants of CAP in individuals with metabolic dysfunction.⁸ However, most existing studies are cross-sectional or primarily focus on adults, with limited evidence from longitudinal Chinese pediatric cohorts. This gap highlights the necessity for prospective pediatric studies to identify potential determinants of CAP and liver fat progression.

Within this context, we established the Wuxi cohort, a longitudinal school-based study initiated in Wuxi in 2023. This cohort was designed to comprehensively evaluate the progression of liver steatosis using CAP, along with detailed assessments of anthropometric characteristics, biochemical markers, dietary intake, and body composition.¹⁴ Consequently, the present study aimed to identify potential determinants of liver steatosis progression in children, thereby providing evidence for early risk stratification and management.

Materials and Methods

Study Design and Population

This study was a prospective cohort study conducted at a primary school in Wuxi, China. The cohort was established in 2023, with baseline data collected from March 12 to March 27, 2023. The design, rationale, and ethical approval for this study have been detailed elsewhere.¹⁴

A total of 1498 participants, aged 6 to 13, were initially enrolled in the cohort study. 221 participants were excluded due to incomplete follow-up, and an additional 82 participants were excluded for missing CAP measurements. The final analysis included a total of 1195 participants (Figure 1).

Figure 1 Flowchart depicting the selection process of study participants. Abbreviation: CAP, controlled attenuation parameter.

Measurement of Biomarkers

In this study, trained professional medical personnel conducted a comprehensive assessment of the subjects, which included a general physical examination (height and weight), CAP measurement, as well as blood, urine, and body composition biomarker assessments. All procedures were performed in accordance with standard operating protocols at the Affiliated Children’s Hospital of Jiangnan University.

Body composition, including 33 biomarkers such as fat-free mass (FFM), protein content, and percent body fat (PBF), was assessed using a bioelectrical impedance analysis device (InBody 370, Seoul, Korea). Participants fasted before the measurement and wore light clothing. During the assessment they held the hand electrodes with their arms relaxed and placed their feet on the footplate electrodes while standing upright.

Blood samples (≥3 mL per tube) and midstream urine samples (10–20 mL) were collected in the early morning after an overnight fast. The samples were stored at 4 °C and analyzed for biochemical parameters using an automated analyzer. Residual specimens were aliquoted and stored at −80 °C for blood and −30 °C for urine for future use.

In addition, multiple systemic immune-inflammatory biomarkers and lipid biomarkers were assessed. The calculation methods for these biomarkers are as follows: inflammatory biomarkers include the systemic immune-inflammation index (SII = platelet count × neutrophil count / lymphocyte count),¹⁵ the systemic inflammation response index (SIRI = monocyte count × neutrophil count / lymphocyte count),¹⁶ the neutrophil-to-lymphocyte ratio (NLR),¹⁶ the neutrophil to albumin ratio (NAR),¹⁷ and the lymphocyte-to-monocyte ratio (LMR).¹⁸ Lipid parameters include the triglyceride-to-high-density lipoprotein cholesterol ratio (THR),¹⁹ the non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio (NHHR),²⁰ and remnant cholesterol, which is calculated as total cholesterol minus low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C).²¹

Assessment of Liver Steatosis

Liver steatosis was assessed by CAP measurements using transient elastography with a FibroScan HANDY^® device (Echosens, Paris, France) equipped with an M probe (3.5 MHz). Two trained physicians measured the participants under conditions of fasting and rest. The participants were positioned supine, with their right arm placed behind their head and their body slightly tilted to the left to expose the 7th to 9th intercostal spaces on the right side. Conductive gel was applied to this area. Valid examinations required at least 10 successful measurements, with a ratio of the interquartile range (IQR) to the median liver stiffness measurement of ≤ 0.30. The median CAP value from the valid measurements was recorded, and the change in CAP was calculated as the difference between the baseline and the 2024 follow-up values. We applied a cut-off value of 248 dB/m for CAP as an indicator of liver steatosis.²² Steatosis progression was defined as the onset of liver steatosis in 2024 among individuals classified as non-steatotic in 2023.

Statistical Analysis

Baseline characteristics of participants who completed assessments in both 2023 and 2024 were compared using a paired t-test. Continuous variables were presented as means ± standard deviations (SD), while categorical variables were expressed as percentages.

Linear and logistic regression analyses were employed to evaluate the associations of biomarkers with (1) changes in CAP and (2) progression of liver steatosis, respectively.

For each biomarker, the robustness of findings was tested using three models: Model A (unadjusted), Model B (adjusted for age, sex, and BMI z-score), and Model C (further adjusted for abdominal circumference, glucose, triglycerides, and HDL-C).

For biomarkers associated with changes in CAP, hierarchical clustering was conducted using the Euclidean distance metric to identify inherent grouping patterns. Additionally, Pearson correlations among these biomarkers were calculated. Subgroup analyses were performed, stratified by sex, BMI, insulin resistance, impaired fasting glucose (IFG), and hypertriglyceridemia, to further evaluate the potential interactions. Restricted cubic spline (RCS) regression, with knots at the 10th, 50th, and 90th percentiles, was employed to assess nonlinear associations.

For the progression of liver steatosis, stepwise logistic regression was employed to identify independent predictors. Discrimination was assessed using receiver operating characteristic (ROC) curve analysis, with the area under the curve (AUC) measured accordingly. Participants with missing values for a specific analysis were excluded from that analysis. A summary of the missing values is presented in Table S1. All statistical analyses were conducted using R (version 4.4.0), considering two-sided P-values of less than 0.05 as statistically significant.

Sensitivity analyses were conducted by repeating the logistic regressions using different CAP cut-off values. We employed pediatric reference values of 225 dB/m, 237 dB/m, and 249 dB/m, as well as the adult cut-off of 275 dB/m recommended by the AASLD guidelines, to assess the consistency of the associations.

Use of Artificial Intelligence Tools

We used ChatGPT and DeepSeek in language correction. The authors reviewed the outputs to ensure accuracy.

Results

Baseline Characteristics

A total of 1195 children participated in both the baseline assessments conducted in 2023 and the follow-up assessments in 2024. The mean age at baseline was 9.22 ± 1.60 years, with a nearly balanced sex distribution of 51.5% boys and 48.5% girls. Compared to the baseline in 2023, the mean CAP significantly decreased in 2024 (193.49 ± 38.81 vs. 188.96 ± 42.95, P = 0.002). Furthermore, several anthropometric and biochemical biomarkers demonstrated statistically significant differences in 2024 compared to baseline (P < 0.05). Among these, the median values for BMI z-score, PBF, direct bilirubin, total protein, aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT), lactate dehydrogenase, glucose, total cholesterol, HDL-C, globulin, lymphocyte percentage, basophil percentage, red blood cell (RBC) count, hematocrit percentage, platelet count, and plateletcrit percentage were significantly reduced, while the remaining parameters showed increases compared to baseline (P < 0.05) (Table 1).

Table 1 Characteristics of Participants (N=1195)

Determinants of Longitudinal CAP Change

The association between changes in CAP and biomarkers is shown in Table 2. In the univariate analysis (Table 2, Model A), changes in CAP were significantly associated with 15 biomarkers, comprising 10 with positive correlations and 5 with negative correlations. Following further adjustments for age, sex, and BMI z-score (Table 2, Model B), 12 biomarkers retained significant associations, including 8 that were positively correlated and 4 that were negatively correlated. In the fully adjusted multivariable model (Table 2, Model C), several biomarkers remained independently associated with longitudinal changes in CAP. The positively associated markers included creatinine (β = 0.97, 95% CI: 0.36–1.59, P = 0.002), neutrophil (β = 0.45, 95% CI: 0.10–0.79, P = 0.012), NLR (β = 4.64, 95% CI: 0.03–9.25, P = 0.049), NAR (β = 0.12, 95% CI: 0.01–0.22, P = 0.029), FFM of the arm (β = 22.40, 95% CI: 3.05–41.76, P = 0.023), FFM of the trunk (β = 3.31, 95% CI: 0.30–6.33, P = 0.031), and neck circumference (β = 4.91, 95% CI: 1.59–8.23, P = 0.004). The negatively associated markers included lymphocyte (β = −0.41, 95% CI: −0.77 to −0.06, P = 0.023), LMR (β = −1.42, 95% CI: −2.65 to −0.20, P = 0.022), and PBF of the arm (β = −0.62, 95% CI: −1.08 to −0.16, P = 0.008).

Table 2 Associations Between Biomarkers and Changes in CAP

As shown in Figure 2, biomarkers clustered into two major groups. The first cluster (green) predominantly characterized by metabolic biomarkers. The second cluster (red) included inflammatory markers (eg, NLR, LMR), FFM, and AST/ALT, reflecting inflammation and lean mass. Notably, NAR (blue) formed a separate cluster, suggesting a distinct correlation pattern.

Figure 2 Clustering and correlation analysis of biomarkers associated with changes in CAP. (A) Hierarchical clustering of biomarkers. (B) Correlation analysis between biomarkers and changes in CAP. Abbreviations: CAP, controlled attenuation parameter; PBF, percent body fat; RDW, red cell distribution width; FFM, fat-free mass; LMR, lymphocyte-to-monocyte ratio; AST/ALT, aspartate aminotransferase - to - alanine aminotransferase ratio; THR, triglyceride-to-high-density lipoprotein cholesterol ratio; NLR, neutrophil-to-lymphocyte ratio; NAR, neutrophil to albumin ratio.

Subgroup Analyses of Association Between CAP Change Value and Biomarkers

Subgroup analyses revealed heterogeneous associations between changes in CAP and the clustered biomarkers. Significant associations with CAP changes were observed among participants without insulin resistance, IFG, or hypertriglyceridemia (P < 0.05). Similar patterns were noted for Cluster 2 biomarkers. No significant interactions were found between changes in CAP and clustered biomarkers across sex, BMI, insulin resistance, IFG, or hypertriglyceridemia strata (Tables S2 and S3). In contrast, significant interactions were found for mean corpuscular hemoglobin concentration (MCHC) and SIRI in sex-stratified analyses, and for PBF and SIRI in IFG stratified analyses (P < 0.05) (Figures S1 and S2).

Non-Linear Association Between CAP Change Value and Biomarkers

As shown in Figure 3, restricted cubic spline analysis revealed nonlinear associations between changes in CAP and lymphocyte, PBF, MCHC, and SIRI. After adjusting for all covariates, the nonlinear associations for PBF and SIRI remained statistically significant (P for nonlinearity < 0.05), while those for lymphocyte and MCHC showed borderline significance (P = 0.052 for lymphocyte and P = 0.057 for MCHC). Specifically, an increase in PBF was associated with a marked reduction in CAP levels of approximately 25%, beyond which the association plateaued. For SIRI, a steep rise in CAP was noted at lower levels, followed by a gradual leveling off, indicating a nonlinear, saturable effect. In contrast, lymphocyte displayed an L-shaped relationship, with CAP decreasing sharply at lower lymphocyte values and stabilizing thereafter. MCHC displayed an inverted U-shaped association with CAP, although this trend did not reach statistical significance after covariate adjustment.

Figure 3 Association between CAP change value and biomarkers in children, assessed using restricted cubic splines. Abbreviations: CAP, controlled attenuation parameter; MCHC, mean corpuscular hemoglobin concentration; PBF, percent body fat; SIRI, systemic inflammation response index.

Independent Predictors of Liver Steatosis Progression

In the follow-up period, 68 children progressed to liver steatosis. After adjusting for all confounding factors (Model C), increased abdominal fat thickness (AFT) was independently associated with steatosis (OR = 12.63; 95% CI: 1.37–129.56; P = 0.029). Additionally, higher neutrophil counts (OR = 1.05; 95% CI: 1.02–1.08; P = 0.002), and reduced lymphocyte counts (OR = 0.96; 95% CI: 0.93–0.99; P = 0.004) also found to be significant predictors. Conversely, However, no significant associations were observed for BFM of the arm and trunk, right thigh fat thickness, or monocyte count following full adjustment (Table 3).

Table 3 Associations Between Biomarkers and Progression to Liver Steatosis

Next, stepwise logistic regression was applied to identify the most robust predictors of steatosis progression. As presented in Table 4, sex (OR = 2.24, 95% CI: 1.21–4.21, P = 0.011), higher BMI z-score (OR = 1.81, 95% CI: 1.18–2.79, P = 0.006), reduced lymphocyte count (OR = 0.96, 95% CI: 0.93–0.99, P = 0.006), and increased BFM of the trunk (OR = 1.19, 95% CI: 1.04–1.38, P = 0.014) were independently associated with progression to steatosis. ROC analysis showed that the model combining sex, BMI z-score, lymphocyte count, and trunk BFM achieved good discrimination for steatosis progression (AUC = 0.803; Figure 4). Among individual predictors, BFM of trunk (AUC = 0.782) and BMI z-score (AUC = 0.746) demonstrated moderate discriminatory ability, while lymphocyte count (AUC = 0.681) and sex (AUC = 0.523) showed relatively weaker performance.

Table 4 Independent Determinants of Liver Steatosis Progression

Figure 4 Receiver operating characteristic (ROC) curves of biomarkers for predicting liver steatosis progression. Abbreviation: BMI, body mass index; BFM, body fat mass.

Sensitivity Analyses

Sensitivity analyses showed that using pediatric reference values (225, 237, 249 dB/m) yielded results consistent with the primary analysis (248 dB/m) for biomarkers (AFT, neutrophil, Lymphocyte) of progression to liver steatosis (all P < 0.05), confirming the robustness of these associations in the pediatric population. In contrast, no significant associations were observed when the adult cut-off of 275 dB/m recommended by the AASLD guidelines was applied (P > 0.05) (Table S4).

Discussion

In this prospective cohort of 1195 Chinese children, we identified several inflammatory and body composition markers associated with longitudinal changes in CAP, and developed a predictive model with good discriminatory performance (AUC = 0.80). Notably, neutrophil-related biomarkers and creatinine were positively associated with steatosis progression, whereas lymphocyte-related biomarkers and peripheral fat distribution showed inverse associations. These findings extend previous cross-sectional studies by providing longitudinal evidence that systemic inflammation and altered body composition play key roles in early liver steatosis progression.

These associations are biologically plausible, given that systemic inflammation has been implicated in the pathogenesis and progression of liver steatosis in previous studies.²³ The excessive accumulation of lipids within hepatocytes initiates localized inflammatory responses and activates immune cells, which release pro-inflammatory cytokines and chemokines, thereby exacerbating systemic inflammation.²⁴ In parallel, signals from adipose tissue further promote the activation and recruitment of liver immune cells, intensifying local inflammation. This inflammatory cascade ultimately contributes to hepatocellular injury and lipid accumulation.²⁵ In adults, numerous studies have proposed inflammatory biomarkers as promising non-invasive indicators of MASLD.^18,26 However, evidence in pediatric populations remains inconclusive and subject to considerable debate. A study in preadolescent children reported that ALT levels were significantly and positively associated with inflammatory markers, including NLR, CRP, and PLR.²⁷ In contrast, another case control study involving children with obesity-related MASLD (N = 267) reported that inflammatory markers including NLR, LMR, and PLR showed no significant associations.²⁸ In addition, emerging evidence suggests that neutrophils are important contributors to the pathogenesis of MASLD.²⁹ By releasing neutrophil elastase,³⁰ myeloperoxidase,³¹ reactive oxygen species, and forming neutrophil extracellular traps,³² neutrophils promote hepatocellular injury, intensify inflammatory responses, and facilitate the development of liver fibrosis.³³ Given the potential mechanistic link between inflammation and liver fat content, this has important implications for the management and treatment of MASLD in children. Currently, lifestyle modification centered on improving dietary quality and increasing physical activity serves as the first-line treatment strategy for these pediatric patients.³⁴ Diet affects inflammation and alters plasma levels of cytokines such as IL-6 and TNF-α.³⁵ The Mediterranean diet is a dietary pattern characterized by antiinflammatory nutrients that include whole grains, legumes, nuts, fish, and low-fat dairy products, as well as healthy oils such as olive oil, vegetables, antioxidant-rich fruits, folic acid, and flavonoids.³⁶ Current evidence from randomized trials in older children demonstrates that the Mediterranean diet reduces hepatic steatosis and improves insulin sensitivity, though such research primarily focuses on secondary prevention.³⁷ Therefore, future strategies for pediatric MASLD may involve developing precise nutritional interventions based on the regulatory relationship between diet and specific inflammatory markers.

Notably, we observed a positive correlation between changes in CAP and creatinine. Similarly, a cross-sectional analysis involving 1387 Chinese children reported that serum creatinine levels were significantly higher in overweight and obese participants, while the uric acid-to-creatinine ratio demonstrated a stronger association with MASLD, showing clear links to hepatic steatosis and liver injury.³⁸ Given that creatinine is almost exclusively excreted by the kidneys, creatinine-related composite biomarkers, such as the urinary C-peptide creatine ratio, have demonstrated significant prognostic value in patients with liver steatosis.³⁹ These findings underscore the importance of careful monitoring of renal function, particularly in pediatric populations.

Furthermore, body composition parameters including arm and trunk FFM, neck circumference, arm PBF, and AFT were significantly associated with liver fat accumulation. These associations may reflect the metabolic effects of regional fat distribution. Upper-body, central, and visceral fat depots have higher lipolytic activity and release larger amounts of free fatty acids (FFA) into the portal circulation.⁴⁰ When the increased portal FFA flux exceeds the liver’s metabolic capacity, triglyceride synthesis rises, leading to hepatic fat deposition.^41,42 Therefore, specific patterns of body composition may independently contribute to liver fat accumulation by influencing lipid mobilization and increasing portal FFA delivery.⁴³ Previous studies have also suggested that body composition plays an important role in liver fat accumulation. For instance, a study involving 227 children showed that lean body mass was positively associated with CAP, whereas lean mass percentage was negatively associated with CAP.⁴⁴ These findings underscore the critical role of early-life adiposity trajectories in the pathogenesis of pediatric MASLD. Interestingly, adult studies have shown a strong association between neck circumference and liver steatosis, proposing optimal neck circumference cut-off values of 40 cm for men and 34 cm for women.⁴⁵ Meanwhile, existing studies have indicated that neck circumference in children can serve as a predictor for clusters of cardiovascular risk factors,⁴⁶ while MASLD is closely linked to childhood obesity and glucolipid metabolic abnormalities. In the future, research could focus on identifying optimal cut-off values in children and promoting the use of neck circumference as a screening tool for liver steatosis. In summary, the body composition parameters identified in this study mainly focused on the upper body region (including the upper limbs, trunk and abdomen). It is suggested that in the early screening of MASLD in children, attention should be paid to the characteristics of segmental fat distribution and individual body size differences. Based on different fat deposition patterns, individualized interventions can be taken. For individuals with excessive fat in the upper limbs and abdomen, protein supplementation⁴⁷ and resistance training⁴⁸ can promote the increase of limb muscle mass and reduce body fat in the upper limbs and abdomen.⁴⁹ High-intensity interval training (HIIT, peak heart rate 80%-90%) is recommended to reduce visceral fat for patients with significantly increased abdominal and visceral fat mass. Ultra-high-intensity interval training (>90% of maximal heart rate) can be considered if reducing whole-body fat burden is the goal.^44,50 However, the above evidence is mainly derived from adult studies, and future studies are needed to verify the efficacy and safety of these intervention strategies in children.

The emergence of NAR as an independent cluster may be attributed to its integration of neutrophil-driven inflammation and albumin-reflecting liver function.⁵¹ Neutrophils contribute to the progression from steatosis to steatohepatitis through ROS production, release of pro-inflammatory factors, and activation of hepatic stellate cells.⁵² Hypoalbuminemia not only indicates impaired hepatic synthetic capacity but also reflects diminished antioxidant and anti-inflammatory capabilities.⁵¹ Although a nonlinear relationship between NAR and hepatic fat content has been established in adults,⁵³ evidence from pediatric cohorts remains limited, warranting further investigation into its potential pathophysiological role.

CAP can serve as a sensitive marker for the detection of early liver steatosis. In the subgroup of metabolically healthy children, elevated CAP values were significantly positively correlated with neck circumference, trunk FFM, neutrophil count, creatinine levels, RDW, and NAR. This finding suggests that inflammatory indicators such as neutrophil count and body composition factors like trunk FFM may directly or indirectly influence the development of hepatic fat deposition, potentially preceding systemic metabolic abnormalities. However, given that this study was based on a school-based natural population cohort, the sample size of children with metabolic abnormalities was limited. Therefore, the interpretation of this association may be subject to selection bias. As such, the findings should be interpreted with caution, and further studies are warranted to elucidate the causal relationship in larger cohorts comprising a greater number of individuals with metabolic abnormalities.

Finally, we developed a prediction model based on independent risk factors for the progression of liver disease. Although several predictive models for pediatric liver steatosis have been proposed, most have been developed within specific subgroups, particularly among obese children.^54,55 However, our model was derived from a single-center cohort and lacks external validation, which may limit its generalizability. Therefore, further studies in larger and more diverse pediatric populations are warranted to validate and refine these findings.

Sensitivity analyses demonstrated robust results using pediatric reference values (225–249 dB/m), whereas no significant association was observed with the adult cut-off (275 dB/m). This discrepancy may be explained by differences in the histopathologic features of MASLD between children and adults. Pediatric MASLD is primarily characterized by zone 1 (periportal) distribution of steatosis, inflammation, and fibrosis, whereas adult MASLD predominantly involves zones 2 and 3, with the most intense changes around the central vein.⁵⁶ Future studies should establish precise CAP thresholds specific to different age groups.

This study is strengthened by its focus on the dynamic assessment of the progression of liver steatosis in a school-aged children. Nonetheless, it is important to recognize that this was an exploratory analysis and is subject to several potential limitations. First, due to the impracticality of performing liver biopsy on a large scale, liver steatosis was assessed using changes in CAP without validation by ALT or abdominal ultrasound, which may have introduced some degree of diagnostic inaccuracy. Second, this study included only a two-year follow-up, potentially limiting the assessment of long-term liver steatosis progression and the stability of biomarker associations. Moreover, the selection of variables may have omitted or failed to include certain relevant factors, raising the possibility of residual confounding that was not accounted for in the models. Finally, the model for predicting steatosis progression was developed based on this specific cohort and lacks external validation due to the unavailability of comparable pediatric datasets with CAP-based quantification of liver fat.

Conclusion

This study evaluated longitudinal changes in liver steatosis among school-aged children using CAP measurements and identified multiple biomarkers associated with the progression of steatosis, predominantly reflecting systemic inflammation and alterations in body composition. These findings underscore the potential utility of these biomarkers in the early identification of liver steatosis in school-aged children. Future research is warranted to validate their predictive performance and to develop stratified management strategies aimed at improving early screening, timely diagnosis, and targeted interventions for children at risk of liver steatosis.