Liver function indicators as risk factors of pediatric metabolic (dysfunction)-associated fatty liver disease: a systematic review and meta-analysis

Abstract

Background

Metabolic (dysfunction)-associated fatty liver disease (MAFLD) is the most common chronic liver disorder in children and adolescents, with its prevalence rising alongside the global childhood obesity epidemic. Liver function indicators offer a potential non-invasive screening alternative, but existing evidence on their association with pediatric MAFLD is inconsistent due to heterogeneous study designs and populations. Therefore, this systematic review and meta-analysis aimed to synthesize global evidence to definitively evaluate liver function indicators as risk factors for MAFLD in children and adolescents.

Methods

Four databases—The Cochrane Library, Embase, Web of Science, and PubMed—were searched from inception to July 5, 2025. The eligible studies were observational in design and focused on children and adolescents (<18 years), comparing MAFLD prevalence/risk between those with abnormal versus normal liver function indicators. Two independent researchers performed literature screening, information collection, and quality evaluation per the eligibility criteria. The ‘meta’ package in R was adopted to compute the odds ratio (OR) and corresponding 95% confidence interval (CI) for the association between liver function indicators and MAFLD. Heterogeneity and publication bias were also assessed.

Results

This meta-analysis incorporated 27 studies, involving 3,237 confirmed MAFLD cases. Levels of alanine aminotransferase (ALT) [OR (95% CI): 1.15 (1.01, 1.30)], gamma-glutamyl transferase (GGT) [1.30 (1.09, 1.56)], and high-density lipoprotein (HDL) [0.97 (0.96, 0.98)] were associated with MAFLD risk. Elevated ALT [OR (95% CI): 20.63 (2.39, 178.09)], total cholesterol (TC) [3.36 (1.15, 9.82)], triglycerides (TG) [4.86 (2.37, 9.98)], low-density lipoprotein (LDL) [3.74 (1.15, 12.19)], and decreased HDL [2.77 (1.97, 3.91)] were identified as potential risk factors for MAFLD in children and adolescents. Overall, subgroup analyses (by confounder adjustment status and study design), sensitivity analyses, and meta-regression did not identify potential sources of heterogeneity. No significant publication bias was observed.

Conclusions

Liver function indicators show promise as screening tools for the early detection of MAFLD susceptibility. This study has several limitations, including a small number of included studies, resulting in heterogeneity, as well as the inherent risk of bias (ROB) in observational designs and the imprecision of some results.

Highlight box.

Key findings

• Levels of alanine aminotransferase (ALT), gamma-glutamyl transferase, and high-density lipoprotein (HDL) were associated with the risk of metabolic (dysfunction)-associated fatty liver disease (MAFLD). Elevated ALT, total cholesterol, triglycerides, low-density lipoprotein, and decreased HDL may be risk factors for MAFLD in children and adolescents.

What is known and what is new?

• Pediatric MAFLD is a growing global health burden, yet, practical and promising screening tools for its early detection remain scarce.

• This review provides synthesized evidence showing that specific liver enzymes and lipid components routinely assessed in clinical settings are strongly associated with MAFLD risk. This confirms their potential in a clinical screening strategy.

What is the implication, and what should change now?

• These findings call for integrating these blood tests into targeted screening of at-risk pediatric populations to enable earlier intervention. Future clinical guidelines should evaluate their adoption to streamline detection and prompt management.

Introduction

Metabolic (dysfunction)-associated fatty liver disease (MAFLD) is the most common chronic liver disorder in children and adolescents and the fastest-growing indication for liver transplantation (1,2). Obesity is a major independent predictor of MAFLD (3). Socioeconomic status, as reflected by the Human Development Index (HDI), is closely associated with obesity incidence and directly influences the disease burden and epidemiological trends of MAFLD (4,5). As childhood obesity rates have risen globally, so as the prevalence of MAFLD, reaching 7.6% in the general pediatric population and 34.2% in obese children (6). MAFLD often has an insidious onset with few early symptoms. However, some patients may develop progressive liver fibrosis, which can lead to cirrhosis or hepatocellular carcinoma (7,8). Additionally, MAFLD is linked to multiple comorbidities, including metabolic syndrome, dyslipidemia, diabetes, hypertension, cardiovascular disease, and renal impairment, which significantly increases the disease burden (9). Therefore, identifying high-risk factors for MAFLD in children and adolescents early on and intervening promptly is crucial to halt disease progression and improve prognosis.

The diagnosis of pediatric MAFLD should be based on evidence of hepatic fat accumulation (steatosis) via histology (biopsy), imaging, or blood biomarkers, along with at least one of the following criteria: overweight or obesity, presence of prediabetes or type 2 diabetes, or evidence of metabolic dysregulation (10). The European Society for Pediatric Gastroenterology, Hepatology, and Nutrition guidelines recommend performing liver function tests and ultrasounds to screen high-risk groups for pediatric MAFLD (11). Ultrasound has become a widely adopted, noninvasive MAFLD screening tool due to its convenience and cost-effectiveness (12). Ultrasound demonstrates good sensitivity and specificity for diagnosing moderate-to-severe hepatic steatosis but is less reliable for detecting mild steatosis (13).

Currently, serum liver function tests are widely used in clinical practice for screening hepatic health due to their convenience and cost-effectiveness (14-16). Liver function indicators serve as biomarkers of liver injury rather than diagnostic markers for MAFLD, though liver injury may precede MAFLD onset. Elevated levels of gamma-glutamyl transferase (GGT), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) typically indicate hepatocyte injury or cholestasis. Alkaline phosphatase (ALP) and total bilirubin levels reflect hepatic metabolic and excretory function, while low-density lipoprotein (LDL) and high-density lipoprotein (HDL) levels demonstrate hepatic lipid metabolism and transport function (14). However, the available research presents conflicting conclusions about the correlation between liver function tests and pediatric MAFLD. This is primarily due to significant heterogeneity in study populations (e.g., age, geography, and ethnicity), differing diagnostic cutoff values, and variability in diagnostic standards. This issue is further complicated by the predominance of small-scale, single-center studies, which lack the statistical power and generalizability of robust, multicenter systematic analyses. These limitations hinder the comprehensive evaluation of liver function tests in predicting pediatric MAFLD risk. Consequently, these limitations challenge clinicians’ ability to accurately assess disease probability based on biochemical markers and hinder the development of effective early intervention strategies.

Therefore, our study uses systematic review and meta-analysis methods to comprehensively integrate global research on the relationship between liver function markers and pediatric MAFLD. Our ultimate goal is to provide high-level evidence for the early detection of high-risk cases and the development of tailored preventive interventions. The findings will advance precision and standardization in the prevention and management of pediatric MAFLD. We present this article in accordance with the PRISMA reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-735/rc).

Methods

This study was registered with the International Prospective Register of Systematic Reviews (PROSPERO, ID: CRD420251105606).

Search strategy

The Cochrane Library, Embase, Web of Science, and PubMed were searched from their inception to July 5, 2025. Reference lists of identified articles were also manually searched for potentially relevant studies. Major search terms used were ‘child’, ‘adolescent’, ‘fatty liver’, ‘risk factors’, ‘association’, ‘odds ratio’, ‘triglycerides’, ‘cholesterol’, ‘alanine transaminase’, ‘aspartate transaminase’, ‘lipoproteins’, and ‘bilirubin’. No restrictions were imposed regarding language or geographic location. The complete search strategy is provided in Appendix 1.

Literature screening

Two researchers (X.F. and F.C.) conducted literature screening independently, and a third reviewer (H.G.) resolved any discrepancies through discussion. Inclusion criteria were: (I) population: children and adolescents (under 18 years old); (II) exposure: abnormal liver function indicator levels; (III) comparators: normal liver function indicator levels; (IV) outcomes: prevalence or risk of MAFLD; and (V) study design: observational studies, including cross-sectional, case-control, and cohort studies. Articles were excluded if they had incomplete data or were abstracts, reviews, comments, or other types of articles.

Information collection

Two reviewers (X.F. and F.C.) extracted the data independently using a pre-developed form. Discrepancies were addressed through consultation with a third reviewer (H.G.). The collected information included the first author, publication year, population, country, study size, mean age or age range of subjects, sex, and body mass index (BMI) of participants, as well as liver function indicators such as ALT, AST, total cholesterol (TC), triglycerides (TG), HDL, LDL, and GGT. All associations between these indicators and MAFLD were considered as primary outcome measures.

Moreover, to evaluate the possible influence of regions with different levels of development, we collected HDI data from the countries in the eligible trials to support further analysis (17).

Quality evaluation

Two reviewers (X.F. and F.C.) independently evaluated the overall quality of the eligible studies using the National Institutes of Health (NIH) Quality Assessment Tool (https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools). They applied the appropriate checklist to each study design (cross-sectional, cohort, or case-control). A third reviewer (H.G.) pooled the results. Each item was scored as ‘yes’, ‘no’, and ‘cannot determine/not reported/not applicable’. Studies that received at least nine or 10 ‘yes’ responses (out of 12 or 14 items, depending on the checklist used) and demonstrated no major methodological concerns were classified as having a low risk of bias (ROB) (Appendix 2).

Statistical analysis

The R software (version 4.5.0) was used for the statistical analysis, primarily employing the ‘meta’ package (version 8.1-0). The pooled effect size was estimated by calculating the hazard ratios (HRs), risk ratios (RRs), or odds ratios (ORs) with 95% confidence intervals (CIs) using either a random-effects model (REM) or a fixed-effects model (FEM), depending on the assessment of heterogeneity. Since liver function indicators can be assessed in two distinct forms—absolute value levels and categorical changes (elevation or decrease) based on clinical cutoff thresholds—which differ fundamentally in clinical relevance and statistical interpretation, we analyzed and merged the results of these two types of studies separately. Heterogeneity across studies was evaluated with the Cochrane Q test and quantified with I² statistics, where I² values exceeding 50% and a P value below 0.05 indicated substantial heterogeneity. We conducted sensitivity analyses and subgroup analyses (stratified by confounder adjustment status and study design) and performed meta-regression (HDI and BMI) to examine possible sources of heterogeneity. Potential publication bias was evaluated visually through funnel plots and statistically via Egger’s linear regression test when the number of eligible studies exceeded ten. If asymmetry was detected, the Trim & Fill method was employed to estimate the number of potentially omitted studies and adjust the pooled effect size accordingly.

Results

Study screening

The initial search yielded 7,179 results. After removing 2,385 duplicates, 4,794 articles were screened based on their titles and abstracts, resulting in the exclusion of 4,741 ineligible articles. Of the remaining 53 articles, 26 were excluded for failing to meet the eligibility criteria: 14 lacked relevant outcomes, 6 were of an ineligible study design, and 6 included adult populations. Consequently, this meta-analysis incorporated 27 articles (Appendices 3,4). Data from Barretto et al.’s [2020] (18) study were excluded due to the log transformation of all values. Data from Sartorio et al.’s [2007] (19) research were omitted because changes in the same outcome direction were opposite. Thus, while these two studies were included in the qualitative review, their data were not incorporated into the meta-analysis. Figure 1 illustrates the literature screening process.

Figure 1.

Literature screening flowchart.

Basic traits and quality evaluation

The meta-analysis included 9,509 participants from 27 studies (20 cross-sectional, five case-control, and two cohort studies) with 3,237 confirmed MAFLD cases. Fourteen of these studies were conducted in Asian countries, including China, Turkey, India, and South Korea. Seven were conducted in European countries, including Italy, Finland, Greece, and Sweden. Three studies were conducted in Africa, one in North America, one in South America, and one in Oceania. Five case-control studies and one cross-sectional study were judged to be of good quality, whereas 19 cross-sectional studies and two cohort studies were considered of fair quality. No studies met the criteria for poor quality. Most studies clearly described the research objectives, study population, and eligibility criteria. However, sample size justification, participant blinding, and repeated exposure assessments were inadequately addressed (Appendix 2). All included articles reported outcome data in the form of ORs. Key characteristics and quality evaluation outcomes are summarized in Appendix 3.

Liver function indicators as risk factors of pediatric MAFLD

ALT

We incorporated 12 studies involving 2,184 subjects to evaluate the association between ALT levels and MAFLD risk. An increase in ALT levels was found to be associated with an increased risk of MAFLD [I²=89.4%, P<0.001; REM; OR (95% CI): 1.15 (1.01, 1.30), P=0.03] (Table 1). Subgroup analysis revealed an increased MAFLD risk with an increase in ALT in unadjusted studies [1.06 (1.04, 1.08)] versus a 20% increase in adjusted studies [1.20 (0.99, 1.45)], with no significant difference between subgroups (P=0.21). Subgroup comparisons by study design showed ORs of 1.04 (95% CI: 1.03, 1.05) for cross-sectional studies, 1.75 (95% CI: 0.66, 4.69) for cohort studies, and 1.32 (95% CI: 0.86, 2.02) for case-control studies. There were no statistically significant differences among these subgroups (P=0.32).

Table 1. Risk factors for MAFLD.

Risk factors	No. of studies	No. of subjects (MAFLD vs. non-MAFLD)	Effect size			Heterogeneity
			OR (95% CI)	P		I2 (%)	P
ALT level	12	2,184 vs. 4,655	1.15 (1.01, 1.30)	0.03		89.4	<0.001
ALT elevation	4	535 vs. 529	20.63 (2.39, 178.09)	0.006		87.1	<0.001
AST level	5	806 vs. 894	1.18 (0.90, 1.55)	0.23		92.7	<0.001
GGT level	4	704 vs. 1740	1.30 (1.09, 1.56)	0.004		91.3	<0.001
TC level	4	542 vs. 519	1.28 (0.92, 1.77)	0.14		89.2	<0.001
TC elevation	5	590 vs. 1,409	3.36 (1.15, 9.82)	0.03		79.4	<0.001
TG level	7	1,084 vs. 1,067	1.16 (0.98, 1.38)	0.08		99.3	<0.001
TG elevation	5	782 vs. 2,074	4.86 (2.37, 9.98)	<0.001		73.8	<0.001
HDL level	7	787 vs. 672	0.97 (0.96, 0.98)	<0.001		45.7	0.06
HDL decrease	6	605 vs. 1,919	2.77 (1.97, 3.91)	<0.001		56.3	0.01
LDL level	4	238 vs. 372	1.54 (0.93, 2.53)	0.09		95.4	<0.001
LDL elevation	5	564 vs. 1,409	3.74 (1.15, 12.19)	0.03		88.8	<0.001

ALT, alanine aminotransferase; AST, aspartate aminotransferase; CI, confidence interval; GGT, gamma-glutamyl transferase; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MAFLD, metabolic (dysfunction)-associated fatty liver disease; No., number; OR, odds ratio; TC, total cholesterol; TG, triglycerides.

Furthermore, 535 subjects from four studies were included to evaluate the association between ALT as a dichotomous variable (normal vs. elevated) and MAFLD risk. The pooled results indicated a statistically significant association between elevated ALT and MAFLD risk [I²=87.1%, P<0.001; REM; OR (95% CI): 20.63 (2.39, 178.09), P=0.006] (Table 1). In subgroup analyses, the pooled OR was 45.53 (95% CI: 0.20, 10,496.17) for unadjusted models and 8.39 (95% CI: 2.01, 35.05) for adjusted models, with no statistically significant difference (P=0.56). Since all of the included studies were cross-sectional, we did not perform a subgroup analysis by study design.

AST

Meta-analysis of five studies including 806 subjects revealed no association between AST levels and MAFLD risk [I²=92.7%, P<0.001; REM; OR (95% CI): 1.18 (0.90, 1.55), P=0.23] (Table 1). Subgroup analysis showed pooled ORs of 1.04 (95% CI: 1.01, 1.07) for unadjusted models and 1.22 (95% CI: 0.87, 1.70) for adjusted models, with no statistically significant difference (P=0.36). However, when the analysis was stratified by study design, a significant difference in the pooled effect estimates was found between cohort studies [OR (95% CI): 2.52 (2.01, 3.16)] and cross-sectional studies [OR (95% CI): 1.04 (1.02, 1.06)] (P<0.001).

GGT

Four studies comprising 704 participants were included to evaluate the link between GGT levels and MAFLD risk. The results demonstrated that an increased GGT level was associated with an elevated MAFLD risk [I²=91.3%, P<0.001; REM; OR (95% CI): 1.30 (1.09, 1.56), P=0.004] (Table 1). Subgroup analyses revealed a pooled OR of 1.22 (95% CI: 1.15, 1.30) for unadjusted studies and 1.33 (95% CI: 1.05, 1.69) for adjusted studies, with no significant difference between subgroups (P=0.49). Stratification by study design revealed pooled ORs of 1.53 (95% CI: 0.97, 2.42) for cohort studies and 1.18 (95% CI: 1.07, 1.30) for cross-sectional studies, with no significant difference between subgroups (P=0.28).

TC

Meta-analysis of four studies comprising 542 subjects revealed no association between TC levels and MAFLD risk [I²=89.2%, P<0.001; REM; OR (95% CI): 1.28 (0.92, 1.77), P=0.14] (Table 1). Subgroup analysis showed pooled ORs of 1.00 (95% CI: 1.00, 1.00) for unadjusted studies and 1.48 (95% CI: 0.91, 2.39) for adjusted studies, with no significant difference between subgroups (P=0.11). Stratification by study design showed pooled ORs of 1.00 (95% CI: 1.00, 1.01) for case-control studies and 1.48 (95% CI: 0.91, 2.39) for cross-sectional studies, with no significant difference between subgroups (P=0.12).

Additionally, a meta-analysis of five studies involving 590 individuals was conducted to evaluate the association of TC as a dichotomous variable (normal vs. elevated) with MAFLD risk. The results revealed a statistically significant association between TC elevation and MAFLD risk [I²=79.4%, P<0.001; REM; OR (95% CI): 3.36 (1.15, 9.82), P=0.03] (Table 1). Subgroup analysis showed pooled ORs of 2.16 (95% CI: 0.42, 11.12) for adjusted studies and 4.03 (95% CI: 0.93, 17.40) for unadjusted studies, with no significant difference between subgroups (P=0.58). When stratified by study design, the pooled OR was 1.35 (95% CI: 0.53, 3.43) for case-control studies and 5.52 (95% CI: 1.17, 26.14) for cross-sectional studies, with no significant difference between subgroups (P=0.13).

TG

Meta-analysis of seven studies (1,084 participants) demonstrated no association between TG levels and MAFLD risk [I²=99.3%, P<0.001; REM; OR (95% CI): 1.16 (0.98, 1.38), P=0.08] (Table 1). Subgroup analyses showed pooled ORs of 1.00 (95% CI: 1.00, 1.00) for unadjusted studies and 1.27 (95% CI: 0.98, 1.66) for adjusted studies, with no significant difference between subgroups (P=0.07). Subgroup comparisons by study design revealed pooled ORs of 1.25 (95% CI: 0.83, 1.87) for case-control studies and 1.00 (95% CI: 1.00, 1.01) for cross-sectional studies, with no significant difference between subgroups (P=0.29).

Furthermore, a meta-analysis of five studies comprising 782 subjects was conducted to evaluate the association of TG as a dichotomous variable (normal vs. elevated) with MAFLD risk. A statistically significant association was found between TG elevation and MAFLD risk [I²=73.8%, P<0.001; REM; OR (95% CI): 4.86 (2.37, 9.98), P<0.001] (Table 1). Subgroup analyses showed that neither study design (P=0.11) nor adjustment for confounding factors (P=0.33) significantly contributed to heterogeneity, and no differences between subgroups were observed.

HDL

A meta-analysis of seven studies involving 787 individuals was conducted to evaluate the link between HDL levels and MAFLD risk. The results indicated that an increase in HDL levels was associated with a lower risk of MAFLD [I²=45.7%, P=0.06; FEM; OR (95% CI): 0.97 (0.96, 0.98), P<0.001] (Table 1). Subgroup analysis showed pooled ORs of 0.97 (95% CI: 0.96, 0.98) for unadjusted studies and 0.98 (95% CI: 0.93, 1.04) for adjusted studies, with no significant difference between subgroups (P=0.65). Stratification by study design showed pooled ORs of 0.98 (95% CI: 0.93, 1.04) for case-control studies and 0.96 (95% CI: 0.95, 0.98) for cross-sectional studies, with no significant difference between subgroups (P=0.54).

Additionally, a meta-analysis of six studies comprising 605 subjects was conducted to evaluate the association between HDL as a dichotomous variable (normal vs. decreased) and MAFLD risk. The results showed a statistically significant association between decreased HDL and MAFLD risk [I²=56.3%, P=0.01; REM; OR (95% CI): 2.77 (1.97, 3.91), P<0.001] (Table 1). Subgroup analyses showed that neither study design (P=0.33) nor adjustment for confounders (P=0.75) significantly contributed to heterogeneity, and no differences between subgroups were observed.

LDL

Four studies comprising 238 participants were included to evaluate the association between LDL levels and MAFLD risk. The meta-analysis revealed no significant association [I²=95.4%, P<0.001; REM; OR (95% CI): 1.54 (0.93, 2.53), P=0.09] (Table 1). Subgroup analysis showed pooled ORs of 0.99 (95% CI: 0.98, 1.00) for unadjusted studies and 1.82 (95% CI: 1.03, 3.21) for adjusted studies, with a significant difference between subgroups (P=0.04). No subgroup analysis by study design was conducted since all of the included studies were cross-sectional.

Furthermore, a meta-analysis of five studies involving 564 subjects was conducted to evaluate the association between LDL as a dichotomous variable (normal vs. elevated) and MAFLD risk. LDL elevation was significantly associated with MAFLD risk [I²=88.8%, P<0.001; REM; OR (95% CI): 3.74 (1.15, 12.19), P=0.03] (Table 1). Subgroup analysis showed pooled ORs of 3.52 (95% CI: 0.92, 13.55) for unadjusted studies and 5.72 (95% CI: 0.14, 230.11) for adjusted studies, with no significant difference between subgroups (P=0.81). Stratification by study design revealed pooled ORs of 1.92 (95% CI: 0.54, 6.82) for case-control studies and 5.33 (95% CI: 0.94, 30.31) for cross-sectional studies, with no significant difference between subgroups (P=0.35).

The forest plots for all liver function indicators are provided in Figures S1-S7. The forest plot for subgroup analysis is summarized in Figure 2.

Figure 2.

Subgroup analysis forest plot of the association of liver function indicators with MAFLD risk. ALT, alanine aminotransferase; AST, aspartate aminotransferase; CI, confidence interval; GGT, gamma-glutamyl transferase; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MAFLD, metabolic (dysfunction)-associated fatty liver disease; OR, odds ratio; TC, total cholesterol; TG, triglycerides.

Sensitivity analysis

A sensitivity analysis was conducted by iteratively removing each study one at a time. The results showed no significant changes in the pooled effect sizes for any of the liver function indicators. This demonstrated the robustness of our findings (Appendix 5).

Publication bias

Funnel plots and Egger’s tests revealed no publication bias for the associations of TG elevation, TG level, and HDL level with MAFLD risk (P>0.05). However, significant publication bias was detected for the association between ALT levels and MAFLD risk (P<0.05). The pooled effect size changed after trim-and-fill adjustment, suggesting that publication bias influenced the results (Figures S8,S9).

Meta-regression

Meta-regression analyses showed no significant association between HDI and the relationship between ALT levels and MAFLD risk (β=0.7523, P=0.64). Similarly, BMI showed no association with the relationship between ALT levels and MAFLD risk (β=0.0011, P=0.84). The rest of liver function indicators had insufficient data to support such an analysis.

Discussion

This meta-analysis included 27 studies, involving 3,237 subjects with MAFLD. The pooled analysis revealed that levels of ALT, GGT, and HDL, elevated ALT, TC, TG, and LDL and decreased HDL, may serve as potential biomarkers for MAFLD risk in children. However, the current evidence base did not reveal statistically significant associations between MAFLD risk and changes in AST, TC, TG, or LDL levels in pediatric populations.

A 2022 meta-analysis by Andreas Vadarlis demonstrated statistically significant differences in ALT, AST, and GGT levels between children with MAFLD and healthy controls (20). Our study verified that ALT level, GGT level, and ALT elevation may serve as risk factors for pediatric MAFLD, while AST levels likely do not. Currently, ALT screening is recommended for MAFLD due to its cost-effectiveness and widespread availability (21). Although ALT screening is advocated as a first-line tool due to its substantial cost advantage over imaging modalities, it has limitations. One key limitation is the lack of consensus on ALT thresholds for defining MAFLD. There is also an undetermined need for sex-, age-, and/or ethnicity-specific cutoffs (16,22-24). For instance, males typically have higher baseline ALT levels than females (16). Additionally, some individuals with MAFLD show ultrasonographic steatosis despite having ALT levels below 40 U/L (25), and ALT levels may not consistently correlate with fibrosis severity (26). ALT, AST, and GGT are fundamental markers of hepatic injury in clinical practice and are part of standard liver function tests. ALT predominantly originates from hepatocytes, while AST comes from the liver, heart, erythrocytes, skeletal muscle, and kidneys (27). This explains why ALT is more specific for hepatic injury. GGT is widely expressed on the cell membranes of the heart, seminal vesicles, kidneys, bile ducts, spleen, and gallbladder (28,29) and has traditionally served as an indicator of hepatic dysfunction, biliary disease, and alcohol consumption (30,31). However, its utility for specific liver conditions may be limited. Consequently, AST and GGT have not been adopted as standalone screening markers for pediatric MAFLD. In cases of elevated ALT, higher AST and GGT levels are associated with worse histology (25). However, elevated AST or GGT levels in the context of normal ALT levels may indicate a condition other than MAFLD.

Our meta-analysis revealed a significant positive association between continuous ALT levels and pediatric MAFLD risk. However, the visual asymmetry in the funnel plot, supported by a statistically significant Egger’s test (P<0.05), suggested the presence of publication bias. The loss of statistical significance of the adjusted effect estimate suggests that the original association may not be robust and could largely be driven by unpublished null findings. This severely undermines confidence in the relationship. Therefore, while an association exists, the true effect of continuous ALT as an MAFLD risk biomarker in the broader pediatric population may be smaller than our initial analysis showed. These findings highlight the need for future large-scale, prospective studies that are designed to minimize publication bias and can provide a more precise estimate of this relationship.

While a 2025 meta-analysis by Xiao et al. identified TG, TC, and LDL levels as risk factors for MAFLD in overweight/obese children and adolescents (32), our study found that only elevated levels of these lipids were risk factors for pediatric MAFLD, not their continuous values. This inconsistency may stem from variations in the study populations and disease stages. Xiao et al.’s study focused on overweight/obese children. Obesity causes systemic chronic inflammation, and high levels of inflammatory factors lead to excessive lipid accumulation in the liver, triggering lipotoxicity and disease progression (33). Normally, the liver stores only small amounts of fatty acids as TG. However, under conditions of nutrient excess and obesity, changes in hepatic fatty acid metabolism typically result in TG accumulation in hepatocytes, leading to MAFLD (34). The group of ‘overweight/obese’ children may generally have elevated lipid levels (such as TC, TG, and LDL). However, our study encompassed a broader pediatric population (including those with normal weight), in which lipid levels exhibited a wider distribution. Here, variable elevation (e.g., defining ‘high’ vs. ‘low’ based on a cutoff value) more effectively distinguishes individuals with already abnormal lipid metabolism and a significantly elevated risk from the ‘normal’ population. In contrast, minor fluctuations within the normal range may be insufficient to yield statistically significant differences in risk for variable levels.

The key strength of our study is that it is the first meta-analysis to systematically evaluate the liver function indicators as risk factors for pediatric MAFLD. However, the current meta-analysis has its limitations. First, there was heterogeneity due to variations in diagnostic criteria (e.g., ultrasound vs. ALT thresholds) and population characteristics (e.g., age, ethnicity, and obesity status) across studies. Furthermore, the small number of eligible studies limited the exploration of potential sources of heterogeneity. Second, some studies failed to adequately adjust for key confounding factors, which may introduce bias. Third, most of the eligible studies were observational in design and carried several known ROB, such as recall bias, selection bias derived from control group selection, and residual confounding. Fourth, the CIs for some pooled effect estimates were notably wide (e.g., OR for elevated ALT =20.63; 95% CI: 2.39, 178.09), indicating substantial imprecision. This is primarily due to the small sample sizes in several of the included studies, which led to unstable estimates. Therefore, the strength of these associations should be interpreted with caution. Finally, while this study suggests a possible association between liver function indicators and the risk of MAFLD in children, caution should be exercised when interpreting these findings. The levels of these indicators are modulated by multiple factors, notably epigenetic programming. The perinatal period is recognized as a critical time for the development of obesity (35) and obesity-related conditions, such as cardiometabolic diseases and MAFLD (36,37). Exposure to maternal overnutrition or undernutrition may increase susceptibility to MAFLD in childhood and accelerate progression to nonalcoholic steatohepatitis throughout one’s lifetime, especially when offspring are exposed postnatally to a high-fat (Western-style) diet (36). Furthermore, maternal obesity and insulin resistance act through the placenta, leading to increased fetal exposure to insulin, lipids, inflammation, and possibly hypoxia. This exposure can program hepatic steatosis before birth (38). Due to the restricted number of studies available, significant differences were only observed in subgroup analyses for a few outcome measures. Therefore, high-quality, prospective studies or matched designs (e.g., matched case-control studies) are needed to validate these associations.

Conclusions

This comprehensive analysis reveals significant correlations between pediatric MAFLD risk and specific liver function indicators, including ALT, GGT, TG, TC, HDL, and LDL. These findings suggest that liver function markers could serve as cost-effective screening tools for the early identification and intervention of pediatric MAFLD. However, due to the limited quantity and quality of the current evidence, further large-scale studies are needed to validate these associations and establish standardized clinical reference values.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.