Abstract
Objective
To compare adipose-inflammatory factor profiles between children with metabolically healthy obesity (MHO) and metabolically unhealthy obesity (MUO), and analyze their associations with non-alcoholic fatty liver disease (NAFLD) severity in MHO.
Methods
This retrospective study included 500 obese children (162 MHO, 338 MUO) and 162 metabolically healthy lean (MHL) controls. Anthropometric, metabolic parameters, and serum levels of key adipose-inflammatory factors (including adiponectin, leptin, resistin, RBP-4, PGRN, TNF-α, IL-6, and CCL2) were compared. ROC curve analysis was used to evaluate diagnostic efficacy of adipose-inflammatory factors for differentiating phenotypes. NAFLD prevalence was assessed, and relationships of adipose-inflammatory factors with NAFLD activity score (NAS) and steatosis, activity, and fibrosis (SAF) score in MHO children with NAFLD were analyzed by Spearman's correlation analysis.
Results
Metabolic parameters and adipose-inflammatory factor levels (leptin, resistin, RBP-4, PGRN, TNF-α, IL-6, CCL2) were significantly higher in MHO than MHL, and further elevated in MUO, while adiponectin showed an inverse trend (all P < 0.05). These factors demonstrated good to excellent diagnostic value for distinguishing MHL from both obese phenotypes (AUC range: 0.695–0.894), and moderate value for distinguishing MHO from MUO (AUC range: 0.636–0.740; all P < 0.001). NAFLD prevalence was 29.01% in MHO vs. 46.15% in MUO (P < 0.001). In MHO children with NAFLD, adiponectin levels correlated negatively with NAS and SAF score (r = −0.668, −0.641), whereas all other factors showed positive correlations (r = 0.468–0.681, all P < 0.001).
Conclusion
MHO children exhibit dysregulation of adipose-inflammatory factors and a considerable risk for NAFLD. These factors, especially adiponectin and leptin, effectively discriminate metabolic phenotypes and correlate with liver injury severity in MHO, suggesting their potential utility as early biomarkers and therapeutic targets.
Keywords: adipose-inflammatory factors, correlation, metabolically healthy obesity, metabolically unhealthy obesity, NAFLD
1. Introduction
Childhood obesity has become one of the most critical public health challenges globally, with its prevalence escalating dramatically in recent decades (1). Obesity is not only an aesthetic concern but also a significant risk factor for various metabolic disorders, often accompanied by insulin resistance (IR), dyslipidemia, hypertension, and non-alcoholic fatty liver disease (NAFLD) (2, 3). However, previous research has revealed substantial metabolic heterogeneity among obese individuals (4). This metabolic heterogeneity has led to a refinement of the obesity concept, classifying it into two phenotypes based on metabolic health: metabolically healthy obesity (MHO) and metabolically unhealthy obesity (MUO). MHO refers to individuals who meet the criteria for obesity but maintain normal metabolic parameters such as blood pressure, blood glucose, and lipids, with relatively preserved insulin sensitivity. In contrast, MUO denotes obesity coexisting with one or more components of metabolic dysregulation (5). The estimated prevalence of MHO among obese adults ranges from 6% to 40%, but reported rates in children and adolescents vary considerably due to the lack of uniform diagnostic criteria (6). Historically, MHO was perceived as benign, and its long-term health risks were not fully recognized. However, growing evidence suggests that MHO may represent a relatively stable or transient metabolic adaptation phase. While the risks of cardiovascular events and type 2 diabetes are lower in individuals with MHO than in those with MUO, these risks remain significantly higher than in their metabolically healthy lean (MHL) counterparts (7). Notably, the risk of NAFLD, one of the most common hepatic complications of obesity, in children with MHO remains unclear. The NAFLD spectrum ranges from simple hepatic steatosis to non-alcoholic steatohepatitis, potentially progressing to fibrosis, cirrhosis, and even hepatocellular carcinoma (8). Disease progression may be more rapid in pediatric NAFLD and is strongly linked to end-stage liver disease and cardiovascular mortality in adulthood (9). Therefore, a deeper investigation into the risk and underlying mechanisms of NAFLD in MHO children is crucial for early identification of high-risk individuals and the implementation of precise interventions.
Obesity-related metabolic disorders are closely associated with a state of chronic low-grade inflammation. Adipose tissue is not merely an energy storage depot but an active endocrine organ secreting various adipokines and inflammatory cytokines, collectively termed adipose-inflammatory factors, which play a central role in regulating energy metabolism, insulin sensitivity, and inflammatory responses. Among these, adiponectin, a protective factor with insulin-sensitizing and anti-inflammatory properties, is typically reduced in obesity (10). Conversely, leptin, resistin, retinol-binding protein 4 (RBP-4), progranulin (PGRN), and classical inflammatory cytokines such as tumor necrosis factor-alpha (TNF-α), interleukin-6 (IL-6), and monocyte chemoattractant protein-1 (CCL2) predominantly exert pro-inflammatory and pro-IR effects, with their levels often elevated in obesity (11–13). However, existing research predominantly focuses on adults, confirming significant associations of these factors with metabolic health and NAFLD progression (14, 15). Currently, systematic studies targeting children with MHO remain scarce. Several critical aspects are yet to be investigated: the expression profiles of key adipose-inflammatory factors, their potential differences compared to MHL children, and whether their dysregulation is linked to the development and progression of NAFLD.
Based on this background, this study aims to comprehensively evaluate the profile of adipose-inflammatory factors in MHO children compared to MHL and MUO children and to analyze the discriminatory efficacy of these factors for different metabolic phenotypes. Furthermore, focusing on the subgroup of MHO children with NAFLD, this study will investigate the correlations between these adipose-inflammatory factors and the severity of histological liver damage, assessed by the NAFLD activity score (NAS) and steatosis, activity, and fibrosis (SAF) score, to elucidate their potential role in liver disease progression in MHO children. The findings may provide important serological evidence for identifying “high-risk MHO” children in the early disease stages and offer theoretical support and potential targets for developing targeted early intervention strategies.
2. Materials and methods
2.1. Demographic characteristics
A retrospective analysis included 500 children with obesity diagnosed at our hospital's pediatric health clinic between January 2023 and July 2025. They were categorized into MHO (n = 162) and MUO (n = 338) groups based on metabolic status. A control group of MHL children (n = 162) was selected during the same period. MHL group: age: [10.00 (8.00, 12.00)] years; gender: 88 males (54.32%) and 74 females (45.68%); ethnicity: 156 Han (96.30%) and 6 ethnic minorities (3.70%); guardian education: primary school 4 (2.47%), junior high 22 (13.58%), senior high 83 (51.23%), and university 53 (32.72%). MHO group: age: [10.00 (8.00, 13.00)] years; gender: 84 males (51.85%) and 78 females (48.15%); ethnicity: 153 Han (94.44%) and 9 ethnic minorities (5.56%); guardian education: primary school 7 (4.32%), junior high 27 (16.67%), senior high 78 (47.15%), and university 50 (30.86%). MUO group: age: [11.00 (8.00, 13.00)] years; gender: 198 males (58.58%) and 140 females (41.42%); ethnicity: 322 Han (95.27%) and 16 ethnic minorities (4.73%); guardian education: primary school 9 (2.66%), junior high 48 (14.20%), senior high 154 (45.56%), and university 127 (37.57%). No statistically significant differences were observed in demographic characteristics among the three groups (P > 0.05), indicating comparability.
2.2. Diagnostic criteria
(1) MHO (5, 16–18): ① Body mass index (BMI) ≥ 95th percentile (for age and sex, applicable for age > 2 years); ② Systolic blood pressure (SBP) and diastolic blood pressure (DBP) < 90th percentile for age and sex; triglycerides (TG) < 1.70 mmol/L; high-density lipoprotein cholesterol (HDL-C) > 1.03 mmol/L; fasting blood glucose (FBG) < 5.6 mmol/L.
(2) MUO (5): Defined as obesity coexisting with at least one component of metabolic abnormality, including elevated blood pressure, dyslipidemia, hyperglycemia, or IR.
(3) NAFLD (19): Chronic hepatic steatosis occurring in children (≤18 years); other causes of hepatic steatosis must be excluded (e.g., alcohol consumption, genetic metabolic diseases, hepatotoxic drug use, infection, or malnutrition). To ensure consistency with the diagnostic criteria and terminology in use during the study enrollment period, the term “NAFLD” was adopted throughout this manuscript. Although the terminology of metabolic dysfunction-associated steatotic liver disease (MASLD) has recently been introduced and is gaining acceptance, this study retained the historical term solely for methodological and temporal consistency.
2.3. Inclusion criteria
① Age of 3–18 years; ② Meeting the diagnostic criteria for MHO, MUO, or NAFLD as defined in section 2.2; ③ First-time diagnosis of obesity at our hospital's pediatric health clinic; ④ Physical condition capable of tolerating study-related examinations and specimen collection procedures; ⑤ Complete clinical data required for the study; ⑥ Legal guardian signed the informed consent, and the study protocol was approved by the Ethics Committee of the Second Affiliated Hospital, Hengyang Medical School, University of South China (No. 2025095).
2.4. Exclusion criteria
① Secondary obesity due to endocrine disorders, genetic syndromes, or medications; ② Presence of liver diseases other than NAFLD that could cause hepatic steatosis or affect liver function, such as active viral hepatitis, autoimmune liver disease, drug-induced liver injury, genetic metabolic liver disease, or cholestatic disorders; ③ Severe systemic diseases, such as significant cardiac or renal disease, malignancy, active autoimmune disease, or acute/chronic infectious period; ④ Recent use (within a specified timeframe prior to the study) of medications potentially significantly affecting core study indicators (body weight, glucose/lipid metabolism, inflammatory status, hepatic fat), such as glucocorticoids, antipsychotics, immunosuppressants, lipid-lowering agents, or hypoglycemic drugs; ⑤ Diagnosed psychiatric illness/cognitive impairment; ⑥ Missing key clinical data or study-required laboratory measurements.
2.5. Observation indicators
(1) Anthropometric indicators: BMI was calculated as weight (kg)/height² (m²). Waist circumference (WC) was measured at the midpoint between the iliac crest and the lower rib margin using a non-elastic tape (precision 0.1 cm).
(2) Metabolic indicators: SBP and DBP were measured three times consecutively using an Omron electronic sphygmomanometer (HEM-7121, Japan), with the average recorded. For the analysis of FBG, TG, and HDL-C, 5 mL of fasting venous blood was collected into a coagulation-promoting tube. After standing at 37°C for 30 min and centrifugation at 3,000 rpm for 10 min (Centrifuge 5424R, Eppendorf, Germany), the serum was analyzed on an automatic biochemical analyzer (Cobas c702, Roche, Switzerland) via enzymatic colorimetric assays. Homeostatic model assessment-IR (HOMA-IR) was calculated as [fasting insulin (μU/mL) × FBG (mmol/L)]/22.5. Fasting insulin was measured by chemiluminescence (ADVIA Centaur XP, Siemens, Germany).
(3) Adipose-inflammatory factors: Adipokines included adiponectin, leptin, resistin, RBP-4, and PGRN. Inflammatory cytokines included TNF-α, IL-6, and CCL2. Fasting venous blood (10 mL) was collected into a coagulation-promoting tube, gently mixed, and centrifuged at 3,000 rpm for 15 min at 4°C. Serum was separated and stored for analysis. Corresponding enzyme-linked immunosorbent assay kits (R&D Systems, Abcam, BioLegend) were used following the manufacturers' protocols. Absorbance was measured using a microplate reader (Epoch2, BioTek, USA), and concentrations were calculated based on standard curves.
(4) NAFLD prevalence: The prevalence of NAFLD among children with different obesity phenotypes (MHO/MUO) was calculated.
(5) NAS (20): Based on liver biopsy histopathology (gold standard), NAS was assessed independently by two senior pathologists in a blinded manner, with arbitration by a third pathologist in case of disagreement. It consists of three components: ① steatosis (S0-3); ② lobular inflammation (LI0-3); ③ hepatocyte ballooning (B0-2). Total score ranges from 0 to 8, with higher scores indicating greater severity.
(6) SAF score (21): Based on liver biopsy histopathology (gold standard), SAF score was assessed independently by two senior pathologists in a blinded manner, with arbitration by a third pathologist in case of disagreement. It consists of three components: ① steatosis (S0-3); ② activity (A0-4), defined as the sum of scores for lobular inflammation (0-2) and hepatocyte ballooning (0–2); ③ fibrosis (F0-4). Total score ranges from 0 to 11, with higher scores indicating greater severity.
2.6. Statistical analysis
SPSS 27.0 was used for statistical analysis, and Prism 8.0.2 for graphing. Categorical data were presented as [n (%)] and analyzed using the χ² test or continuity-corrected χ² test. Normally distributed quantitative data were expressed as mean ± standard deviation ( ± s), with group comparisons performed using independent samples t-tests. Non-normally distributed quantitative data were expressed as M (P25, P75), with group comparisons performed using the Mann–Whitney U-test. Receiver operating characteristic (ROC) curves were used to evaluate the diagnostic efficacy of adipose-inflammatory factors in distinguishing MHL from MHO, MHL from MUO, and MHO from MUO. The Youden's index was calculated as sensitivity + specificity − 1. Spearman's correlation analysis was used to examine the relationship between adipose-inflammatory factor levels and NAS/SAF scores in MHO children with NAFLD. The significance level was set at P < 0.05.
3. Results
3.1. Comparison of anthropometric and metabolic indicators
BMI, WC, SBP, DBP, FBG, TG, and HOMA-IR were significantly higher in both the MHO and MUO groups compared to the MHL group, with all parameters higher in the MUO group than in the MHO group. HDL-C was lower in both MHO and MUO groups compared to the MHL group and was lower in the MUO group than in the MHO group (P < 0.05) (Table 1).
Table 1.
Comparison of anthropometric and metabolic indicators among the three groups.
| Indicator | MHL group (n = 162) | MHO group (n = 162) | MUO group (n = 338) |
|---|---|---|---|
| BMI (kg/m2) | 17.65 (16.39, 19.12) | 22.52 (20.80, 24.78)a | 25.52 (23.35, 27.15)a,b |
| WC (cm) | 63.00 (59.00, 68.00) | 84.50 (79.00, 91.00)a | 92.00 (85.00, 99.00)a,b |
| SBP (mmHg) | 99.00 (97.00, 102.00) | 109.00 (105.00, 112.00)a | 118.00 (115.00, 121.00)a,b |
| DBP (mmHg) | 64.00 (63.00, 65.00) | 72.00 (68.00, 74.00)a | 78.00 (76.00, 80.00)a,b |
| FBG (mmol/L) | 4.82 (4.62, 5.02) | 5.18 (5.02, 5.34)a | 5.82 (5.48, 6.13)a,b |
| TG (mmol/L) | 0.88 (0.76, 1.00) | 1.22 (1.05, 1.41)a | 2.14 (1.78, 2.43)a,b |
| HDL-C (mmol/L) | 1.50 (1.36, 1.64) | 1.21 (1.13, 1.29)a | 0.95 (0.83, 1.08)a,b |
| HOMA-IR | 1.89 (1.51, 2.23) | 2.94 (2.39, 3.51)a | 5.13 (4.14, 6.03)a,b |
BMI, Body mass index; WC, Waist circumference; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; FBG, Fasting blood glucose; TG, Triglycerides; HDL-C, High-density lipoprotein cholesterol; HOMA-IR, Homeostatic model assessment-insulin resistance.
P < 0.05 compared to the MHL group.
P < 0.05 compared to the MHO group.
3.2. Comparison of adipose-inflammatory factors
Leptin, resistin, RBP-4, PGRN, TNF-α, IL-6, and CCL2 levels were significantly higher in both MHO and MUO groups compared to the MHL group, with all levels higher in the MUO group than in the MHO group. Adiponectin levels were significantly lower in both MHO and MUO groups compared to the MHL group and were lower in the MUO group than in the MHO group (P < 0.05) (Table 2).
Table 2.
Comparison of adipose-inflammatory factors among the three groups.
| Factor | MHL group (n = 162) | MHO group (n = 162) | MUO group (n = 338) |
|---|---|---|---|
| Adiponectin (μg/mL) | 11.93 (9.77, 13.73) | 9.18 (7.95, 10.61)a | 8.34 (7.17, 9.39)a,b |
| Leptin (ng/mL) | 5.80 (4.61, 7.20) | 8.82 (6.40, 11.22)a | 12.60 (9.06, 15.69)a,b |
| Resistin (ng/mL) | 8.36 (6.84, 10.01) | 11.06 (8.89, 13.32)a | 13.11 (10.01, 15.97)a,b |
| RBP-4 (μg/mL) | 32.96 (27.48, 37.78) | 39.61 (33.98, 45.79)a | 46.25 (38.78, 53.36)a,b |
| PGRN (ng/mL) | 45.96 ± 10.40 | 58.62 ± 13.49a | 74.87 ± 20.03a,b |
| TNF-α (pg/mL) | 6.93 (5.67, 8.20) | 8.34 (6.77, 10.13)a | 9.57 (7.67, 11.76)a,b |
| IL-6 (pg/mL) | 2.25 (1.76, 2.71) | 2.92 (2.24, 3.65)a | 3.74 (2.48, 4.93)a,b |
| CCL2 (pg/mL) | 187.13 ± 42.91 | 229.50 ± 57.80a | 279.29 ± 83.94a,b |
RBP-4, Retinol-binding protein 4; PGRN, Progranulin; TNF-α, Tumor necrosis factor-alpha; IL-6, Interleukin-6; CCL2, Monocyte chemoattractant protein-1.
P < 0.05 compared to the MHL group.
P < 0.05 compared to the MHO group.
3.3. Diagnostic efficacy of adipose-inflammatory factors in distinguishing MHL from MHO
When differentiating MHL from MHO, all adipose-inflammatory factors showed significant discriminatory ability. The AUC (95% CI) values were 0.787 (0.738–0.836) for adiponectin, 0.770 (0.718–0.822) for leptin, 0.751 (0.698–0.804) for resistin, 0.746 (0.693–0.798) for RBP-4, 0.763 (0.712–0.814) for PGRN, 0.695 (0.639–0.752) for TNF-α, 0.705 (0.648–0.762) for IL-6, and 0.714 (0.658–0.770) for CCL2 (all P < 0.001) (Table 3, Figure 1).
Table 3.
Diagnostic efficacy of adipose-inflammatory factors in distinguishing MHL from MHO.
| Factor | AUC | 95% CI | Youden's index | Sensitivity | Specificity | P |
|---|---|---|---|---|---|---|
| Adiponectin | 0.787 | 0.738–0.836 | 0.470 | 87.70% | 59.30% | <0.001 |
| Leptin | 0.770 | 0.718–0.822 | 0.488 | 53.70% | 95.10% | <0.001 |
| Resistin | 0.751 | 0.698–0.804 | 0.414 | 53.10% | 88.30% | <0.001 |
| RBP-4 | 0.746 | 0.693–0.798 | 0.377 | 55.60% | 82.10% | <0.001 |
| PGRN | 0.763 | 0.712–0.814 | 0.420 | 55.60% | 86.40% | <0.001 |
| TNF-α | 0.695 | 0.639–0.752 | 0.309 | 43.20% | 87.70% | <0.001 |
| IL-6 | 0.705 | 0.648–0.762 | 0.370 | 50.00% | 87.00% | <0.001 |
| CCL2 | 0.714 | 0.658–0.770 | 0.352 | 46.90% | 88.30% | <0.001 |
Figure 1.
ROC curves of adipose-inflammatory factors for distinguishing MHL from MHO. ROC, Receiver operating characteristic; MHO, Metabolically healthy obesity; MUO, Metabolically unhealthy obesity.
3.4. Diagnostic efficacy of adipose-inflammatory factors in distinguishing MHL from MUO
Similarly, these adipose-inflammatory factors also demonstrated significant diagnostic performance in distinguishing MHL from MUO. The AUC (95% CI) values were 0.877 (0.842–0.912) for adiponectin, 0.894 (0.867–0.922) for leptin, 0.826 (0.791–0.861) for resistin, 0.859 (0.828–0.890) for RBP-4, 0.891 (0.863–0.918) for PGRN, 0.793 (0.755–0.832) for TNF-α, 0.772 (0.732–0.812) for IL-6, and 0.822 (0.787–0.858) for CCL2 (all P < 0.001) (Table 4, Figure 2).
Table 4.
Diagnostic efficacy of adipose-inflammatory factors in distinguishing MHL from MUO.
| Factor | AUC | 95% CI | Youden's index | Sensitivity | Specificity | P |
|---|---|---|---|---|---|---|
| Adiponectin | 0.877 | 0.842–0.912 | 0.611 | 93.20% | 67.90% | <0.001 |
| Leptin | 0.894 | 0.867–0.922 | 0.723 | 77.20% | 95.10% | <0.001 |
| Resistin | 0.826 | 0.791–0.861 | 0.565 | 62.70% | 93.80% | <0.001 |
| RBP-4 | 0.859 | 0.828–0.890 | 0.599 | 72.20% | 87.70% | <0.001 |
| PGRN | 0.891 | 0.863–0.918 | 0.684 | 74.60% | 93.80% | <0.001 |
| TNF-α | 0.793 | 0.755–0.832 | 0.474 | 53.00% | 94.40% | <0.001 |
| IL-6 | 0.772 | 0.732–0.812 | 0.550 | 61.80% | 93.20% | <0.001 |
| CCL2 | 0.822 | 0.787–0.858 | 0.581 | 63.00% | 95.10% | <0.001 |
Figure 2.
ROC curves of adipose-inflammatory factors for distinguishing MHL from MUO.
3.5. Diagnostic efficacy of adipose-inflammatory factors in distinguishing MHO from MUO
For the discrimination between MHO and MUO, all adipose-inflammatory factors remained statistically significant, although their overall diagnostic performance was relatively modest. The AUC (95% CI) values for distinguishing MHO from MUO were 0.656 (0.604–0.708) for adiponectin, 0.734 (0.692–0.777) for leptin, 0.644 (0.596–0.692) for resistin, 0.684 (0.638–0.731) for RBP-4, 0.740 (0.698–0.783) for PGRN, 0.636 (0.586–0.685) for TNF-α, 0.648 (0.601–0.695) for IL-6, and 0.681 (0.635–0.728) for CCL2 (all P < 0.001) (Table 5, Figure 3).
Table 5.
Diagnostic efficacy of adipose-inflammatory factors in distinguishing MHO from MUO.
| Factor | AUC | 95% CI | Youden's index | Sensitivity | Specificity | P |
|---|---|---|---|---|---|---|
| Adiponectin | 0.656 | 0.604–0.708 | 0.250 | 86.70% | 38.30% | <0.001 |
| Leptin | 0.734 | 0.692–0.777 | 0.398 | 49.10% | 90.70% | <0.001 |
| Resistin | 0.644 | 0.596–0.692 | 0.271 | 38.80% | 88.30% | <0.001 |
| RBP-4 | 0.684 | 0.638–0.731 | 0.300 | 42.30% | 87.70% | <0.001 |
| PGRN | 0.740 | 0.698–0.783 | 0.413 | 53.60% | 87.70% | <0.001 |
| TNF-α | 0.636 | 0.586–0.685 | 0.241 | 34.60% | 89.50% | <0.001 |
| IL-6 | 0.648 | 0.601–0.695 | 0.331 | 39.30% | 93.80% | <0.001 |
| CCL2 | 0.681 | 0.635–0.728 | 0.339 | 45.00% | 88.90% | <0.001 |
Figure 3.
ROC curves of adipose-inflammatory factors for distinguishing MHO from MUO.
3.6. Prevalence of NAFLD among children with different obesity phenotypes
Among the 162 MHO children, 47 (29.01%) had NAFLD, while 115 (70.99%) did not. Among the 338 MUO children, 156 (46.15%) had NAFLD, while 182 (53.85%) did not. The difference between the two groups was statistically significant (χ² = 13.343, P < 0.001).
3.7. Correlation between adipose-inflammatory factors and NAS in MHO children with NAFLD
Spearman's correlation analysis in the 47 MHO children with NAFLD showed a negative correlation between adiponectin and NAS (r = −0.668, P < 0.001). Leptin, resistin, RBP-4, PGRN, TNF-α, IL-6, and CCL2 all showed positive correlations with NAS (r = 0.572, 0.484, 0.468, 0.548, 0.633, 0.624, 0.562, all P < 0.001) (Figure 4).
Figure 4.
Scatter plots showing correlations between adipose-inflammatory factors and NAS. (a), Adiponectin and NAS; (b), Leptin and NAS; (c), Resistin and NAS; (d), RBP-4 and NAS; (e), PGRN and NAS; (f), TNF-α and NAS; (g), IL-6 and NAS; h, CCL2 and NAS. NAS, NAFLD activity score; RBP-4, Retinol-binding protein 4; PGRN, Progranulin; TNF-α, Tumor necrosis factor-alpha; IL-6, Interleukin-6; CCL2, Monocyte chemoattractant protein-1.
3.8. Correlation between adipose-inflammatory factors and SAF score in MHO children with NAFLD
Spearman's correlation analysis in the 47 MHO children with NAFLD showed a negative correlation between adiponectin and SAF score (r = −0.641, P < 0.001). Leptin, resistin, RBP-4, PGRN, TNF-α, IL-6, and CCL2 all showed positive correlations with SAF score (r = 0.681, 0.560, 0.518, 0.676, 0.508, 0.532, 0.641, all P < 0.001) (Figure 5).
Figure 5.
Scatter plots showing correlations between adipose-inflammatory factors and SAF score.
4. Discussion
Research on the metabolic heterogeneity of childhood obesity has progressed from phenotypic description to molecular mechanisms. Adipose-inflammatory factors, serving as a critical bridge connecting energy metabolism and immune regulation, have become a recent focus due to their differential expression patterns across obesity subtypes. With advances in clinical diagnostics, the detection rate of NAFLD in the pediatric population has been increasing annually (up to 36.1%), but consensus on its pathogenesis in MHO children remains elusive (22). Existing evidence suggests that individuals with MHO, despite normal traditional metabolic parameters, may experience covert remodeling of the adipose tissue microenvironment, manifesting as an imbalance in adipokine secretion and early activation of a low-grade inflammatory state. This subclinical alteration may be a potential driver of hepatic steatosis (23). Furthermore, compared to adults, dynamic changes in adipose-inflammatory factors during childhood are more susceptible to influences from developmental stage, dietary patterns, and genetic background, possibly resulting in unique patterns of association with NAFLD progression. Therefore, elucidating the aberrant profile of adipose-inflammatory factors in MHO children and their intrinsic link to liver injury holds significant theoretical value for redefining the concept of metabolic health in childhood obesity and refining the NAFLD risk stratification system.
The results of this study indicated that although MHO children did not exhibit significant glucose or lipid metabolic abnormalities (blood pressure, blood glucose, lipids, etc., although slightly higher than the MHL group, remain within normal ranges), their circulating levels of various adipose-inflammatory factors were already markedly altered. This was characterized by significantly reduced adiponectin levels, alongside elevated levels of leptin, resistin, RBP-4, PGRN, and the inflammatory cytokines TNF-α, IL-6, and CCL2. Furthermore, these levels lay intermediate between those of the MHL and MUO groups. This finding partially aligns with studies in adult MHO populations (24), suggesting that adipose tissue dysfunction may precede abnormalities in traditional metabolic parameters in MHO children. This may be because adipocytes in obese children have relatively strong proliferative and differentiation capacities. Adipose tissue might temporarily maintain metabolic stability through compensatory expansion. However, when long-term energy intake exceeds expenditure, excessive adipocyte recruitment during this compensatory process triggers a series of subclinical changes: First, relative hypoxia in adipose tissue due to inadequate oxygen supply activates hypoxia-inducible factor-1α, promoting the expression and release of pro-inflammatory cytokines such as TNF-α and IL-6 (25). Second, enhanced endoplasmic reticulum stress in adipocytes, via the PERK/eIF2α pathway, inhibits the synthesis and secretion of adiponectin, a key factor with insulin-sensitizing and anti-inflammatory effects, while upregulating the transcription of the pro-inflammatory factor leptin (26). Third, increased macrophage infiltration (particularly M1 pro-inflammatory macrophages) in adipose tissue, secreting chemokines like CCL2, forms an inflammatory amplification loop, further exacerbating the imbalance of adipose-inflammatory factors (27).
ROC curve analysis in this study further confirmed the core value of adipose-inflammatory factors in distinguishing different metabolic phenotypes. In distinguishing MHL from MHO, adiponectin (AUC = 0.787) and leptin (AUC = 0.770) demonstrated the highest diagnostic efficacy, with Youden's indices of 0.470 and 0.488, respectively, suggesting their potential as sensitive indicators for identifying early metabolic abnormality. Leptin, in particular, emerged as the best single factor for distinguishing MHL from MHO with a specificity of 95.10%. This may be closely related to its central role in energy metabolism regulation. Elevated leptin levels not only reflect increased fat mass but also indicate the early onset of central leptin resistance, a key transition point from metabolic health to metabolic abnormality (28). In distinguishing MHL from MUO, leptin (AUC = 0.894) and PGRN (AUC = 0.891) showed the most prominent diagnostic efficacy, with AUC values approaching 0.90, indicating that when obesity progresses to the metabolically unhealthy stage, changes in these two factors are highly specific. In contrast, traditional inflammatory cytokines like TNF-α (AUC = 0.695) showed lower efficacy in distinguishing MHO from MHL, possibly due to their smaller magnitude of change in early obesity, not yet reaching the threshold sufficient for phenotype discrimination. Notably, even in distinguishing the metabolically more similar MHO from MUO, leptin (AUC = 0.734) and PGRN (AUC = 0.740) maintained moderate-to-good diagnostic value, outperforming other factors. This result suggests that leptin and PGRN may be involved in key regulatory processes underlying the differentiation of metabolic phenotypes in obesity, and their level changes could serve as potential markers for predicting the transition from MHO to MUO.
Regarding NAFLD risk, this study found a prevalence of 29.01% in MHO children. Although lower than the 46.15% in the MUO group, this rate is significantly higher than previously reported rates in normal-weight children (approximately 7.6%) (29). These data suggest that even with normal metabolic parameters, obesity itself places children at a significant risk of liver injury. More importantly, in MHO children with NAFLD, levels of adipose-inflammatory factors showed moderate-to-strong correlations (|r| = 0.468–0.681) with the severity of histological liver damage (NAS, SAF score), indicating that these factors are not only involved in the onset of NAFLD but are also significantly associated with disease progression. This aligns with the adipose-liver axis theory in adult studies, where adipose tissue-derived factors directly affect the liver via the portal circulation, regulating lipid deposition, insulin signaling, and inflammation (30, 31). For example, adiponectin can inhibit hepatic lipogenesis and enhance fatty acid oxidation by activating the AMPK pathway; its decline weakens hepatic protection. Leptin can promote liver inflammation and fibrosis via the JAK/STAT pathway. Inflammatory factors such as resistin and TNF-α exacerbate IR and hepatocyte damage through the NF-κB pathway (32–34). This study validates the significant correlations of these factors with NAS and SAF score in MHO children, suggesting that even in the absence of overt metabolic abnormalities, dysregulation of adipose-inflammatory factors may independently drive the onset and progression of NAFLD. This finding deepens the re-evaluation of the benign MHO phenotype: its metabolic health may only represent a state of systemic metabolic compensation, while target organs (like the liver) may already be subject to potential damage.
In summary, children with MHO exhibit dysregulation of adipose-inflammatory factors and an increased risk of NAFLD. Adipose-inflammatory factors, particularly adiponectin and leptin, effectively distinguish metabolic phenotypes of obesity and demonstrate moderate-to-strong correlations with the severity of liver injury in MHO children, suggesting their potential value as targets for early intervention. Compared to previous studies focusing on MUO populations, this study emphasizes that MHO children also warrant attention regarding liver health risks, and adipose-inflammatory factors could serve as early warning indicators and intervention targets. Notably, the histological findings based on liver biopsy not only reveal the molecular characteristics of liver injury in MHO children but also provide crucial biological evidence for advancing the transition of pediatric clinical practice toward non-invasive diagnostic strategies. Currently, imaging techniques represented by shear wave elastography (SWE) and ultrasound-based fat quantification have demonstrated potential for dynamic monitoring as alternatives to biopsy (35, 36). Integrating the key factors identified in this study with these non-invasive imaging parameters holds promise for constructing a multidimensional comprehensive assessment model. This approach could not only provide a non-invasive and precise tool for stratifying liver injury risk in MHO children but also open new avenues for dynamically monitoring disease trajectories and evaluating intervention efficacy. Based on this rationale, future research could conduct longitudinal cohort studies to dynamically track the evolution of adipokine profiles in MHO children and, in conjunction with non-invasive imaging techniques, investigate the critical turning points in progression to MUO and NAFLD. This would provide evidence-based foundations for developing early intervention strategies targeting adipose-inflammatory factors.
Funding Statement
The author(s) declared that financial support was not received for this work and/or its publication.
Footnotes
Edited by: Andrea Cassidy-Bushrow, Henry Ford Health System, United States
Reviewed by: Maximilian Zeyda, Medical University of Vienna, Austria
Nurullah DAĞ, İnönü University, Türkiye
Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.
Ethics statement
The studies involving humans were approved by the Second Affiliated Hospital, Hengyang Medical School, University of South China (No. 2025095). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants' legal guardians/next of kin.
Author contributions
JL: Methodology, Resources, Supervision, Writing – review & editing. HS: Data curation, Formal analysis, Writing – original draft. QX: Validation, Visualization, Writing – review & editing. PZ: Resources, Supervision, Writing – review & editing. XL: Conceptualization, Investigation, Project administration, Writing – original draft.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
- 1.The Lancet Diabetes E. Childhood obesity: a growing pandemic. Lancet Diabetes Endocrinol. (2022) 10(1):1. 10.1016/S2213-8587(21)00314-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Salama M, Balagopal B, Fennoy I, Kumar S. Childhood obesity, diabetes. And cardiovascular disease risk. J Clin Endocrinol Metab. (2023) 108(12):3051–66. 10.1210/clinem/dgad361 [DOI] [PubMed] [Google Scholar]
- 3.Chung ST, Krenek A, Magge SN. Childhood obesity and cardiovascular disease risk. Curr Atheroscler Rep. (2023) 25(7):405–15. 10.1007/s11883-023-01111-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.He D, Qiu Y, Yan M, Zhou T, Cheng Z, Li J, et al. Associations of metabolic heterogeneity of obesity with frailty progression: results from two prospective cohorts. J Cachexia Sarcopenia Muscle. (2023) 14(1):632–41. 10.1002/jcsm.13169 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Maternal and Child Health Research Society PCoOCiWaC, Chinese Expert Committee on the Definition and Management of Metabolically Healthy Obesity in Children. The expert consensus on definition of metabolically healthy obesity and screening metabolically healthy obesity in Chinese children. Chin J Woman Child Health Res. (2019) 30(12):1487–90. 10.3969/j.issn.1673-5293.2019.12.001 [DOI] [Google Scholar]
- 6.Prince RL, Kuk JL, Ambler KA, Dhaliwal J, Ball GD. Predictors of metabolically healthy obesity in children. Diabetes Care. (2014) 37(5):1462–8. 10.2337/dc13-1697 [DOI] [PubMed] [Google Scholar]
- 7.Mayoral LP, Andrade GM, Mayoral EP, Huerta TH, Canseco SP, Rodal Canales FJ, et al. Obesity subtypes, related biomarkers & heterogeneity. Indian J Med Res. (2020) 151(1):11–21. 10.4103/ijmr.IJMR_1768_17 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Nassir F. NAFLD: mechanisms, treatments, and biomarkers. Biomolecules. (2022) 12(6):824. 10.3390/biom12060824 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zhou X, Lin X, Chen J, Pu J, Wu W, Wu Z, et al. Clinical spectrum transition and prediction model of nonalcoholic fatty liver disease in children with obesity. Front Endocrinol. (2022) 13:986841. 10.3389/fendo.2022.986841 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Engin A. Adiponectin resistance in obesity: adiponectin leptin/insulin interaction. Adv Exp Med Biol. (2024) 1460:431–62. 10.1007/978-3-031-63657-8_15 [DOI] [PubMed] [Google Scholar]
- 11.Liu L, Shi Z, Ji X, Zhang W, Luan J, Zahr T, et al. Adipokines, adiposity, and atherosclerosis. Cell Mol Life Sci. (2022) 79(5):272. 10.1007/s00018-022-04286-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Seliem EM, Azab ME, Ismail RS, Nafeaa AA, Alotaibi BS, Negm WA. Green coffee bean extract normalize obesity-induced alterations of metabolic parameters in rats by upregulating adiponectin and GLUT4 levels and reducing RBP-4 and HOMA-IR. Life. (2022) 12(5):693. 10.3390/life12050693 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wu Y, Ma Y. CCL2-CCR2 signaling axis in obesity and metabolic diseases. J Cell Physiol. (2024) 239(4):e31192. 10.1002/jcp.31192 [DOI] [PubMed] [Google Scholar]
- 14.An SM, Cho SH, Yoon JC. Adipose tissue and metabolic health. Diabetes Metab J. (2023) 47(5):595–611. 10.4093/dmj.2023.0011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Schuster S, Cabrera D, Arrese M, Feldstein AE. Triggering and resolution of inflammation in NASH. Nat Rev Gastroenterol Hepatol. (2018) 15(6):349–64. 10.1038/s41575-018-0009-6 [DOI] [PubMed] [Google Scholar]
- 16.Li H, Ji C, Zong X, Zhang Y. Height and weight standardized growth charts for Chinese children and adolescents aged 0 to 18 years. Chin J Pediatr. (2009) 47(7):487–92. 10.3760/cma.j.issn.0578-1310.2009.07.003 [DOI] [PubMed] [Google Scholar]
- 17.Li H, Ji C, Zong X, Zhang Y. Body mass index growth curves for Chinese children and adolescents aged 0 to 18 years. Chin J Pediatr. (2009) 47(7):493–8. 10.3760/cma.j.issn.0578-1310.2009.07.004 [DOI] [PubMed] [Google Scholar]
- 18.Dong Y, Ma J, Song Y, Dong B, Wang Z, Yang Z, et al. National blood pressure reference for Chinese Han children and adolescents aged 7 to 17 Years. Hypertension. (2017) 70(5):897–906. 10.1161/HYPERTENSIONAHA.117.09983 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Eslam M, Alkhouri N, Vajro P, Baumann U, Weiss R, Socha P, et al. Defining paediatric metabolic (dysfunction)-associated fatty liver disease: an international expert consensus statement. Lancet Gastroenterol Hepatol. (2021) 6(10):864–73. 10.1016/S2468-1253(21)00183-7 [DOI] [PubMed] [Google Scholar]
- 20.Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology. (2005) 41(6):1313–21. 10.1002/hep.20701 [DOI] [PubMed] [Google Scholar]
- 21.European Association for the Study of the L, European Association for the Study of D and European Association for the Study of O. EASL-EASD-EASO clinical practice guidelines for the management of non-alcoholic fatty liver disease. Diabetologia. (2016) 59(6):1121–40. 10.1007/s00125-016-3902-y [DOI] [PubMed] [Google Scholar]
- 22.Shaunak M, Byrne CD, Davis N, Afolabi P, Faust SN, Davies JH. Non-alcoholic fatty liver disease and childhood obesity. Arch Dis Child. (2021) 106(1):3–8. 10.1136/archdischild-2019-318063 [DOI] [PubMed] [Google Scholar]
- 23.Lonardo A, Mantovani A, Lugari S, Targher G. Epidemiology and pathophysiology of the association between NAFLD and metabolically healthy or metabolically unhealthy obesity. Ann Hepatol. (2020) 19(4):359–66. 10.1016/j.aohep.2020.03.001 [DOI] [PubMed] [Google Scholar]
- 24.Aisike G, Kuerbanjiang M, Muheyati D, Zaibibuli K, Lv MX, Han J. Correlation analysis of obesity phenotypes with leptin and adiponectin. Sci Rep. (2023) 13(1):17718. 10.1038/s41598-023-43550-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hofmann E, Soppert J, Ruhl T, Gousopoulos E, Gerra S, Storti G, et al. The role of macrophage migration inhibitory factor in adipose-derived stem cells under hypoxia. Front Physiol. (2021) 12:638448. 10.3389/fphys.2021.638448 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Vasu K, Ramachandiran I, Chechi A, Khan K, Khan D, Kaufman R, et al. Translational control of murine adiponectin expression by an upstream open reading frame element. RNA Biol. (2023) 20(1):737–49. 10.1080/15476286.2023.2256094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lang Y, Liu Y, Ye C, Tang X, Cheng Z, Xie L, et al. Loss of LEAP-2 alleviates obesity-induced myocardial injury by regulating macrophage polarization. Exp Cell Res. (2023) 430(2):113702. 10.1016/j.yexcr.2023.113702 [DOI] [PubMed] [Google Scholar]
- 28.Abaturov A, Nikulina A. Obesity in children with leptin receptor gene polymorphisms. Acta Medica (Hradec Kralove). (2021) 64(3):158–64. 10.14712/18059694.2021.27 [DOI] [PubMed] [Google Scholar]
- 29.Anderson EL, Howe LD, Jones HE, Higgins JP, Lawlor DA, Fraser A. The prevalence of non-alcoholic fatty liver disease in children and adolescents: a systematic review and meta-analysis. PLoS One. (2015) 10(10):e0140908. 10.1371/journal.pone.0140908 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Buzzetti E, Pinzani M, Tsochatzis EA. The multiple-hit pathogenesis of non-alcoholic fatty liver disease (NAFLD). Metab Clin Exp. (2016) 65(8):1038–48. 10.1016/j.metabol.2015.12.012 [DOI] [PubMed] [Google Scholar]
- 31.Bence KK, Birnbaum MJ. Metabolic drivers of non-alcoholic fatty liver disease. Mol Metab. (2021) 50:101143. 10.1016/j.molmet.2020.101143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Fang H, Judd RL. Adiponectin regulation and function. Compr Physiol. (2018) 8(3):1031–63. 10.1002/cphy.c170046 [DOI] [PubMed] [Google Scholar]
- 33.Zhu Y, Mao H, Peng G, Zeng Q, Wei Q, Ruan J, et al. Effect of JAK-STAT pathway in regulation of fatty liver hemorrhagic syndrome in chickens. Anim Biosci. (2021) 34(1):143–53. 10.5713/ajas.19.0874 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Engin A. Reappraisal of adipose tissue inflammation in obesity. Adv Exp Med Biol. (2024) 1460:297–327. 10.1007/978-3-031-63657-8_10 [DOI] [PubMed] [Google Scholar]
- 35.Akcicek M, Dag N. Evaluation of hepatic steatosis in obese children and adolescents using immune-inflammatory markers and shear wave elastography. J Ultrason. (2025) 25(100):1–6. 10.15557/jou.2025.0001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Dag N, Sarici B, Igci G, Yagin FH, Yilmaz S, Kutlu R. Diagnostic performance of ultrasound-based liver fat quantification with reference to magnetic resonance imaging proton density fat fraction and histology. J Clin Ultrasound. (2025) 53(6):1265–71. 10.1002/jcu.24010 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.