Reclassifying Pediatric NAFLD Using the Steatotic Liver Disease Framework: A Multicenter Retrospective Study from the NASH CRN

Abstract

Background & Aims:

The terminology for hepatic steatosis and nonalcoholic fatty liver disease (NAFLD) was revised under the umbrella of steatotic liver disease (SLD), with metabolic dysfunction-associated steatotic liver disease (MASLD) as the primary subtype. MASLD is defined by hepatic steatosis plus at least one cardiometabolic risk factor (CMRF). A new category, Met-ALD, describes MASLD with alcohol consumption below the defined thresholds for alcohol-associated liver disease (ALD). While adult studies have demonstrated strong concordance between NAFLD and MASLD, the applicability of this framework in children remains unclear.

Approach & Results:

We assessed children clinically diagnosed with NAFLD and enrolled in the NASH CRN who had available liver histology. Clinical and demographic data, including body mass index (BMI), hepatotoxic medication use, and alcohol intake, were analyzed. Liver biopsies were centrally reviewed to confirm hepatic steatosis and evaluate for alternative etiologies. Participants were reclassified using the SLD framework. Among 1,019 children diagnosed with NAFLD, 858 (84%) met MASLD criteria. The average number of CMRFs per participant was 2.7 ± 1.1; 41 (4.7%) met all five. Thirty-three participants (3.2%) were reclassified as Met-ALD, a prevalence that rose to 5.4% among adolescents. Sixty-six children (6.5%) were reclassified as drug-induced SLD.

Conclusions:

Most children with NAFLD met MASLD criteria, but nearly 1 in 6 were reclassified based on alcohol use or medication exposure. These findings highlight the need for a systematic diagnostic approach accounting for metabolic risk factors, alcohol use, and medication-related liver injury.

INTRODUCTION

Nonalcoholic fatty liver disease (NAFLD) was historically defined as hepatic steatosis in individuals without significant alcohol consumption or other known liver diseases.¹ This diagnostic category encompassed a spectrum of hepatic injury, from isolated steatosis to progressive inflammation and fibrosis, and has been associated with increased risks for cirrhosis, cardiometabolic complications, and even premature mortality in youth.^2,3 Over the past two decades, NAFLD emerged as the most common chronic liver disease in children, affecting up to 10% of the pediatric population.^4,5 However, the reliance on exclusionary diagnostic criteria, insufficient emphasis on underlying pathophysiology, and potential for stigma have long been points of contention.⁶ These limitations prompted the recent paradigm shift toward a more inclusive and pathophysiologically grounded nomenclature.

Under this updated nomenclature, steatotic liver disease (SLD) is defined by the presence of hepatic steatosis from any cause, and serves as an umbrella category encompassing multiple subtypes. Metabolic dysfunction–associated steatotic liver disease (MASLD) replaces NAFLD as the primary diagnosis in individuals with hepatic steatosis and at least one cardiometabolic risk factor (e.g., obesity, dysglycemia, dyslipidemia, or hypertension). Additional diagnostic subtypes include Met-ALD, which refers to individuals who meet MASLD criteria and consume alcohol below the defined thresholds for alcohol-associated liver disease (ALD), and cryptogenic SLD, which applies to individuals with steatosis lacking identifiable metabolic or secondary causes. SLD resulting from alternative etiologies—such as drug-induced liver injury, monogenic conditions, or infections—is classified separately. Importantly, the MASLD framework is based on attribution to associated factors rather than exclusion of alternatives, allowing for classification across overlapping or coexisting risk profiles.⁶

While adult studies have shown that MASLD captures most individuals previously classified as NAFLD, concordance varies by cohort characteristics, diagnostic approach, and metabolic risk profiles.^7–10 In biopsy-confirmed registries, concordance often exceeds 99%, but is lower in population-based studies, particularly those with leaner populations or incomplete exclusion of secondary causes.^11,12 Importantly, pediatric populations have been largely absent from this literature, despite unique genetic, developmental, and environmental factors that may influence steatosis. Heritable monogenic disorders, evolving metabolic physiology, and comparatively limited cumulative alcohol exposure create a clinical landscape that differs from adults and may affect the sensitivity and specificity of MASLD criteria in youth. Rigorous evaluation of this framework in pediatrics is therefore essential for accurate diagnosis and appropriate classification.

To address this gap, we analyzed a large, multicenter cohort of children with biopsy-confirmed NAFLD enrolled in the Nonalcoholic Steatohepatitis Clinical Research Network (NASH CRN). At the time of enrollment, diagnoses were made based on clinical judgment and histologic findings, consistent with prevailing pediatric standards. For the present study, we applied the MASLD framework retrospectively, using centrally reviewed liver histology and systematically collected data on cardiometabolic risk factors (CMRF), medication exposure, and alcohol use. Our goals were to determine the proportion of children who meet MASLD criteria, to assess the frequency and characteristics of reclassified subgroups based on factors such as alcohol and medications, and to explore how histologic and genetic findings differ across diagnostic categories. In doing so, we aimed to clarify both the utility and limitations of the MASLD framework when applied to pediatric steatotic liver disease.

METHODS

Study Design and Population

This cross-sectional study analyzed data from three pediatric cohorts within the NASH CRN to evaluate the applicability of the updated MASLD terminology in children diagnosed with NAFLD by pediatric gastroenterologists in their clinical practice at a NASH CRN clinical center. These clinical diagnoses were based on liver histology and other clinical information available at the time of diagnosis and enrollment. For the present study, we used central review of those same liver biopsies to confirm or exclude hepatic steatosis and systematically apply the SLD classification framework. Participants were under 18 years old and enrolled in one of the following NASH CRN cohorts between 2004 and 2024: Database (2004–2009), Database 2 (NCT01061684; 2010–2019), or Database 3 (NCT04454463; 2020–2024). The core inclusion and exclusion criteria, enrollment procedures, and data collection methods for each phase of these cohorts have been previously described.^13,14 Study protocols required enrollment within 90 days of clinical liver biopsy. Children without available histology were excluded. Ethical approval was obtained from the Institutional Review Boards of all participating institutions. All research was conducted in accordance with both the Declarations of Helsinki and Istanbul. Written informed consent was secured from parents or guardians, and assent was obtained from children aged 8 years or older. The MASLD framework was applied retrospectively to these clinically diagnosed NAFLD cases, enabling classification based on contemporary criteria while preserving the integrity of original enrollment decisions.

Medical History and Physical Examination

Comprehensive demographic and clinical data were collected during standardized study visits. Demographic variables included age, sex, race, and ethnicity. Medical histories were documented using structured interviews to capture comorbid conditions and medication use. Specific attention was paid to medications targeting cardiometabolic conditions (e.g., antihypertensives, antidiabetic agents, lipid-lowering therapies) and hepatotoxic drugs (e.g., corticosteroids, 2nd generation antipsychotics, anticonvulsants, and methotrexate) to evaluate their potential contribution to drug-induced steatotic liver disease (DISLD). Use of medications with known or suspected hepatosteatogenic potential did not automatically preclude enrollment in the NASH CRN studies. At the time, many such medications particularly those associated with weight gain or insulin resistance were not considered incompatible with a diagnosis of NAFLD when supported by liver histology and clinical context. Rather than reflecting competing etiologies, these exposures were understood to potentially coexist with or contribute to NAFLD, consistent with the multifactorial understanding of the disease. In the current analysis, these cases were reclassified as DISLD based on the MASLD framework, which emphasizes associations rather than exclusive causality.

For participants aged 12 years and older, alcohol consumption was assessed using validated tools, including the Alcohol Use Disorders Identification Test (AUDIT) and its abbreviated version, the Alcohol Use Disorders Identification Test - Consumption (AUDIT-C).¹⁵ At the time these studies were conducted, the primary purpose of alcohol assessment was to determine whether participants met the criteria for NAFLD. The AUDIT and AUDIT-C were chosen because they quantify both the amount and pattern of alcohol intake, allowing for application of these criteria. Although the MASLD nomenclature was introduced after data collection, these instruments are also well-suited to our current aims, as they enable differentiation of MASLD from Met-ALD and ALD. Both the American Academy of Pediatrics and the National Institute on Alcohol Abuse and Alcoholism support their use in adolescents ≥12 years of age, and prior studies have demonstrated their validity in this population.^16–18

Anthropometric measurements, including weight and height, were used to calculate body mass index (BMI) and BMI z-scores, adjusted for age and sex per the 2022 CDC extended BMI-for-growth standards. Severe obesity was defined as a BMI ≥35 or ≥120% of the 95th percentile. Blood pressure was measured in a seated position using a calibrated sphygmomanometer and an appropriately sized cuff, with at least two consecutive measurements averaged for accuracy. Age-height-sex blood pressure percentiles were computed.¹⁹

Laboratory Evaluation

Participants underwent fasting (minimum 12 hours) prior to phlebotomy. Blood samples were analyzed for hepatic, metabolic, and hematologic parameters. Hepatic tests included alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT), total bilirubin, direct bilirubin, and albumin. Metabolic assessments encompassed lipid panels (total cholesterol, high-density lipoprotein [HDL], low-density lipoprotein [LDL], triglycerides) and glycemic markers (fasting glucose, hemoglobin A1c [HbA1c]). Patatin-like phospholipase domain containing 3 (PNPLA3) genotyping (rs738409) was performed on available samples to assess genetic predisposition to hepatic steatosis and to evaluate biological differences among diagnostic subgroups.

Liver Histology

For this analysis, liver histology reviewed by the NASH CRN Pathology Committee served as the definitive confirmation of hepatic steatosis. Although participants had been clinically diagnosed with NAFLD at enrollment, only those with ≥5% macrovesicular steatosis on central pathology review were included in the SLD classification. The original clinical diagnoses were based on liver histology and supporting clinical data available at the time, including exclusion of other causes of liver disease by history, laboratory evaluation, and local pathology assessment.

Biopsies were stained with hematoxylin–eosin and Masson’s trichrome and centrally reviewed by the NASH CRN Pathology Committee according to the Network Histology Scoring Manual, as originally described by Kleiner et al. Reviews were conducted at quarterly committee meetings attended by a quorum of the nine hepatopathologists in the committee. Glass slides were examined on a multiheaded microscope, with reviewers blinded to whether the biopsy originated from a pediatric or adult study, the study center, and all clinical and demographic data. All histologic features were scored by consensus. Macrovesicular steatosis was quantified as the percentage of hepatocytes affected and graded: grade 0 (<5%), grade 1 (5%–33%), grade 2 (34%–66%), and grade 3 (>66%). Lobular inflammation was categorized by the number of inflammatory foci per 200x field (none, <2, 2–4, or >4), while portal inflammation was graded as none, mild, or more than mild. Hepatocyte ballooning was described as none, few, or many. Fibrosis was staged: stage 0 (no fibrosis), stage 1a (mild perisinusoidal requiring trichrome), stage 1b (moderate perisinusoidal fibrosis), stage 1c (portal/periportal fibrosis only), stage 2 (perisinusoidal and periportal fibrosis), stage 3 (bridging fibrosis), and stage 4 (cirrhosis).

Biopsies with <5% macrovesicular steatosis were excluded from SLD classification. The presence of small droplet steatosis alone, while consistent with SLD, did not meet criteria for NAFLD due to the absence of macrovesicular fat. Biopsies with ≥5% macrovesicular steatosis were subcategorized as NAFLD not NASH, borderline zone 1 NASH, borderline zone 3 NASH, or definite NASH (details in Supplemental Table 1). Features suggestive of alternative liver diseases—such as drug-induced liver injury, autoimmune hepatitis, alcohol-related injury, or viral hepatitis—were also documented. Final biopsy classifications included NAFLD subtypes, small droplet steatosis (SLD but not NAFLD), evidence of other liver diseases, or insufficient steatosis (<5%).

Applying the SLD Framework

Our primary objective was to apply the MASLD framework to a real-world clinical cohort of children diagnosed with NAFLD. The following SLD diagnostic framework and definitions were therefore applied to all participants with a diagnosis of NAFLD and available liver histology (see Figure 1).

FIGURE 1:

Participant flow diagram designating selection of children enrolled in NASH CRN longitudinal studies that are classified as MASLD. Abbreviations: ALD, alcohol-associated liver disease; DISLD, drug-induced steatotic liver disease; MASLD, metabolic dysfunction–associated steatotic liver disease; Met-ALD, Metabolic dysfunction and alcohol-associated steatotic liver disease; NASH CRN, Nonalcoholic Steatohepatitis Clinical Research Network; SLD, steatotic liver disease.

A diagnosis of MASLD required the presence of macrovesicular steatosis and at least one CMRF. Qualifying CMRFs included: (1) BMI >85th percentile for age and sex; (2) dysglycemia, defined as fasting glucose ≥100 mg/dL, hemoglobin A1c ≥5.7%, or a diagnosis of type 2 diabetes; (3) elevated blood pressure, defined as systolic or diastolic pressure >95th percentile for age, or ≥130/85 mmHg for adolescents, or current use of antihypertensive medication; (4) elevated triglycerides (>100 mg/dL for children <10 years or >150 mg/dL for those ≥10 years, or use of lipid-lowering medication); and (5) low HDL cholesterol (<40 mg/dL).⁶

Participants with MASLD who reported any alcohol consumption below thresholds for ALD (<14 units/week for males and <7 units/week for females) were conservatively reclassified as having Met-ALD. This approach is supported by the AASLD Pediatric MASLD Clinical Practice Statement, which affirms that no level of alcohol use is considered safe in adolescents, and is consistent with the Rinella et al. Delphi consensus, which recognized that pediatric-specific thresholds were not established and would require separate consideration. Those who exceeded these thresholds were classified as ALD.

DISLD classification required the presence of macrovesicular steatosis on central pathology review and documented exposure, at the time of biopsy, to a medication with known or suspected steatogenic potential (Supplementary Table 2). This included glucocorticoids, second-generation antipsychotics, certain anticonvulsants, methotrexate, and other less commonly prescribed agents, identified from peer-reviewed literature and pharmacology references. Classification was based on association rather than proven causality, consistent with the MASLD framework, and could include participants with coexisting CMRFs. In all cases, medication exposure preceded liver biopsy, but causality was not formally assessed due to the retrospective design.

Monogenic SLD encompassed cases of steatosis attributable to specific genetic disorders. Cryptogenic SLD was reserved for participants with steatosis in the absence of metabolic dysfunction or identifiable secondary etiologies.⁶

Data analysis

Descriptive statistics were used to characterize baseline demographics, clinical features, and histological findings. Continuous variables were expressed as means (±standard deviations), and categorical variables as counts (percentages). Group comparisons used Student’s t-tests or ANCOVA for continuous data and chi-squared tests for categorical data. Fisher’s exact test was employed for small expected cell sizes. Highly skewed distributions, such as liver enzyme levels, were compared using the Kruskal-Wallis test. The prevalence of each diagnostic subgroup was illustrated with pie charts, and Venn diagrams were used to show CMRF distributions. A subgroup analysis restricted to participants aged ≥12 years was performed to assess potential confounding by age in comparisons involving Met-ALD. To assess site-level heterogeneity, we fit a mixed-effects logistic regression with clinic site as a random intercept and study phase as a fixed effect (outcome: MASLD vs. non-MASLD), reporting the variance estimate, intraclass correlation coefficient (ICC), and likelihood ratio test for the random effect.

All statistical analyses were conducted using SAS (v9.4, SAS Institute, Cary, NC) and Stata/IC (v16.1, StatCorp LLC, College Station, TX).

RESULTS

Study population

A total of 1,019 children with a clinical diagnosis of NAFLD and available liver histology were included (Table 1). The mean age was 13.0 ± 2.8 years, with males comprising 71% of the cohort. With regards to ethnicity and race, 72% identified as Hispanic and 63% as White (including Hispanic White). Obesity was highly prevalent: mean BMI was 32.2 ± 6.4 kg/m², mean BMI z-score was 2.9 ± 1.4, and 60% had severe obesity. The mean ALT was 111 ± 103 U/L, and the mean GGT was 46 ± 41 U/L.

Table 1.

Characteristics of Study Participants

Participant characteristic	Total N=1019	MASLD N=858	NAFLD not MASLD N=161	P-value
Age, yr, mean (SD)	13.0 (2.8)	12.8 (2.8)	14.3 (2.6)	<0.001
Sex, N (%)				0.88
Male	723 (71)	608 (71)	115 (71%)
Female	296 (29)	250 (29)	46 (29)
Ethnicity, N (%)				<0.001
Hispanic	735 (72)	654 (76)	85 (53%)
Anthropometrics
Weight, kg, mean (SD)	83.1 (25.9)	81.5 (25.1)	91.2 (28.4)	<0.001
Height, cm, mean (SD)	1.59 (0.14)	1.58 (0.14)	1.64 (0.14)	<0.001
BMI, mean (SD)	32.2 (6.4)	32.0 (6.2)	33.3 (7.2)	0.02
BMI Z-score, mean (SD)	2.9 (1.4)	3.0 (1.3)	2.9 (1.6)	0.68
Systolic blood pressure, mm Hg, mean (SD)	120 (13)	120 (13)	121 (12)	0.66
Diastolic blood pressure, mm Hg, mean (SD)	68 (9)	68 (9)	68 (10)	0.83
Laboratory Values
Aspartate aminotransferase, U/L, mean (SD)	64 (54)	65 (55)	60 (48)	0.35
Alanine aminotransferase, U/L, mean (SD)	111 (103)	113 (104)	102 (96)	0.43
Alkaline phosphatase, U/L, mean (SD)	224 (105)	230 (103)	188 (109.7)	<0.001
γ-glutamyl transpeptidase, U/L, mean (SD)	46 (41)	45 (37)	52 (59)	0.07
Total bilirubin, mg/dL, mean (SD)	0.51 (0.29)	0.50 (0.28)	0.55 (0.33)	0.06
Platelet count, 103 cells/mcl, mean (SD)	293 (70)	296 (70)	276 (68)	0.001
Triglycerides, mg/dL, mean (SD)	148 (80)	147 (76)	155 (97)	0.21
Total cholesterol, mg/dL, mean (SD)	164 (36)	164 (36)	165 (37)	0.77
HDL, mg/dL, mean (SD)	39 (9)	39 (9)	40 (11)	0.71
LDL, mg/dL, mean (SD)	96 (30)	97 (30)	96 (31)	0.72
Glucose, mg/dL, mean (SD)	91 (22)	92 (22)	89 (24)	0.28
HbA1c, %, mean (SD)	5.5 (0.8)	5.5 (0.8)	5.5 (0.9)	0.98
PNPLA3 genotype, N (%)				<0.001
CC	107 (15)	79 (13)	28 (23)
CG	252 (34)	192 (31)	60 (49)
GG	377 (51)	343 (56)	34 (28)
Histology
Definite Steatohepatitis, N (%)	190 (19)	162 (19)	28 (24)	0.20
Fibrosis Stages 2–4	268 (26)	233 (27)	35 (22)	0.15
Ballooning, Grade 2, N (%)	86 (8)	70 (8)	16 (10)	0.46

This table presents characteristics of children diagnosed with NAFLD by NASH CRN investigators. Participants are grouped by MASLD classification status. P values reflect comparisons between participants reclassified as MASLD and those with NAFLD reclassified as something other than MASLD. Continuous variables were compared using t-tests; categorical variables using chi-square or Fisher’s exact tests, as appropriate. Liver enzyme values were compared using Kruskal-Wallis tests due to skewed distributions. Genotype data were unavailable for 283 participants.

Histology

Central pathology review identified 48 children (4.7%) with <5% steatosis, excluding them from SLD classification. Among the remaining participants, alternative histologic findings were observed in 12 children: 9 with isolated small droplet steatosis, 1 with viral hepatitis, and 2 with overt histologic features of alcohol-related liver injury. Four additional biopsies showed features consistent with autoimmune hepatitis (AIH) in the context of SLD (Figure 1).

Medical History Review

Of the 959 participants with confirmed SLD and no histologic exclusion criteria for MASLD, 66 (6.9%) were taking medications known to induce hepatic steatosis and were reclassified as having DISLD. One additional participant was diagnosed with autoimmune hepatitis based on clinical and laboratory findings (Figure 1).

Alcohol Intake

Alcohol use was reported by 32 participants (3.1%). One child met thresholds for ALD and was classified accordingly. The remaining 31 reported consumption below ALD thresholds and were reclassified as Met-ALD (Figure 1).

Physical Examination and Laboratory Evaluation

Among the 861 participants without DISLD or ALD, the prevalence of individual CMRFs was as follows: overweight or obesity in 848 (98%), elevated blood pressure in 286 (33%), dysglycemia in 290 (34%), elevated triglycerides in 373 (43%), and low HDL cholesterol in 498 (58%). Three participants (0.3%) had no CMRFs or alternative etiologies and were classified as cryptogenic SLD.

Diagnostic Reclassification

The initial cohort of 1,019 children with a clinical diagnosis of NAFLD was systematically reclassified using the updated SLD framework (Figure 2). Forty-eight children (4.7%) were excluded due to insufficient steatosis. Of the remaining participants, 858 (84.2%) met MASLD criteria, including five (0.5%) who also had AIH. Thirty-one children (3.0% of the total population) were reclassified as Met-ALD, one (0.1%) as ALD, 66 (6.5%) as DISLD, and three (0.3%) as cryptogenic SLD (Figure 2). In a mixed-effects logistic regression, 6% of the variance in classification outcomes (MASLD vs. non-MASLD) was attributable to site-level differences (ICC 0.06; 95% CI 0.02–0.19), representing modest but statistically significant heterogeneity (likelihood ratio test p = 0.002).

FIGURE 2:

Reclassification of children with NAFLD using the SLD framework. This pie chart illustrates the distribution of diagnostic categories among children initially diagnosed with NAFLD following reclassification under the updated steatotic liver disease (SLD) framework. The majority met criteria for metabolic dysfunction–associated steatotic liver disease (MASLD), while others were reclassified based on alcohol use (Met-ALD, ALD), exposure to hepatotoxic medications (DISLD), or absence of metabolic risk factors (cryptogenic SLD). A subset was excluded from SLD classification due to insufficient hepatic steatosis or alternative etiologies, including microvesicular steatosis (n=9) and viral hepatitis (n = 1). Percentages represent the proportion of the total cohort assigned to each diagnostic category. Abbreviations: ALD, alcohol-associated liver disease; DISLD, drug-induced steatotic liver disease; MASLD, metabolic dysfunction–associated steatotic liver disease; Met-ALD, Metabolic dysfunction and alcohol-associated steatotic liver disease; SLD, steatotic liver disease.

Comparisons Between Diagnostic Groups

MASLD Subgroup

Children with MASLD comprised the majority of the cohort (84.2%) and had demographic and clinical characteristics similar to the overall study population. Among participants with available genotyping, the MASLD group had the highest prevalence of the PNPLA3 GG genotype (56%). The distribution of CMRFs among MASLD participants is shown in Figure 3. The average number of CMRFs per participant was 2.7 ± 1.1. The most common pattern was elevated BMI alone (123 children, 14.3%), followed by elevated BMI with low HDL cholesterol (114 children, 13.3%), and elevated BMI combined with both low HDL and high triglycerides (102 children, 11.9%). Notably, 41 children (4.7%) exhibited all five CMRFs. This high-risk subgroup had a significantly greater prevalence of definite steatohepatitis (41% vs. 18%, p < 0.001) and a higher frequency of clinically significant fibrosis (stages 2–4: 43% vs. 26%, p < 0.001) compared to MASLD participants with fewer than five risk factors.

FIGURE 3:

Distribution of cardiometabolic risk factors among children with MASLD. This Venn diagram illustrates the overlap and distribution of five cardiometabolic risk factors among children diagnosed with MASLD. BMI (elevated BMI: ≥85th percentile for age and sex), Glucose (dysglycemia: fasting glucose ≥100 mg/dL, HbA1c ≥5.7%, or diagnosis of type 2 diabetes), Blood Pressure (elevated blood pressure: ≥95th percentile for age or ≥130/80 mmHg if ≥13 years old), HDL (low HDL cholesterol: <40 mg/dL), and Triglycerides (elevated triglycerides: >100 mg/dL for children <10 years, >150 mg/dL for children ≥10 years) are shown as overlapping regions. Each number within the diagram represents the count of participants exhibiting the corresponding combination of risk factors. The central intersection indicates participants with all five cardiometabolic risk factors, while peripheral regions reflect those with fewer or isolated risk factors. Abbreviation: BMI, body mass index.

Met-ALD Subgroup

Among adolescents aged ≥12 years with available alcohol intake data, 5.4% (33 of 612) were reclassified as Met-ALD. Compared to their peers with MASLD, adolescents with Met-ALD had higher weight (109.2 ± 27.6 vs. 94.2 ± 22.7 kg; p < 0.001), and higher BMI (37.4 ± 8.0 vs. 33.9 ± 6.3 kg/m²; p = 0.002) and BMI Z-score (3.3 ± 1.7 vs. 2.8 ± 1.2; p = 0.03). GGT levels were significantly higher in the Met-ALD group (64 vs. 46 U/L; p = 0.02), while alkaline phosphatase levels were lower (132 ± 81 vs. 179 ± 93 U/L; p < 0.001). Hispanic ethnicity was less common among Met-ALD participants (52% vs. 70%; p = 0.02). There were no significant differences in sex distribution, cardiovascular parameters, liver enzymes, hematologic indices, lipid levels, or glycemic markers. The PNPLA3 GG genotype was significantly less common in adolescents with Met-ALD (31%) compared to those with MASLD (49%; p = 0.03), suggesting distinct underlying risk profiles.

DISLD Subgroup

Children with DISLD tended to be younger (mean age, 13.5 ± 3.0 years) and had a higher mean BMI z-score (3.3 ± 1.7) compared to those with MASLD, though these differences did not reach statistical significance (p = 0.06). Hypertriglyceridemia (59%) and low HDL (73%) were significantly more common in the DISLD group compared to MASLD (43% and 58%, respectively; both p < 0.002). Psychiatric or neurological comorbidities were also more frequent in children with DISLD (p < 0.002). The PNPLA3 GG genotype was least common in this group (22%; p < 0.001 vs. MASLD), further supporting the biological distinction of DISLD as a reclassified category.

DISCUSSION

In this large, multicenter cohort of children with clinically diagnosed NAFLD including liver histology, we evaluated the applicability of the updated SLD framework. Although all participants were diagnosed with NAFLD by pediatric gastroenterologists using clinical and histologic data available at enrollment, reclassification using centralized pathology review and standardized criteria revealed that 84.2% met MASLD criteria. Nearly 16% were reclassified as having Met-ALD, DISLD, or cryptogenic SLD, highlighting the importance of structured and comprehensive diagnostic evaluation for children with hepatic steatosis.

Among children classified as MASLD, most had multiple CMRFs. Those with all five qualifying risk factors exhibited a significantly greater burden of liver disease, including higher rates of steatohepatitis and clinically significant fibrosis. These findings align with adult studies linking fibrosis progression to cumulative metabolic burden.^20,21 For example, in the Framingham Heart Study, liver stiffness measured by vibration-controlled transient elastography (VCTE) was independently associated with multiple CMRFs, even after adjusting for BMI and steatosis, reinforcing the role of systemic metabolic dysfunction in liver disease progression.²² These findings highlight the importance of comprehensive cardiometabolic screening in children with MASLD, as well as the need for early intervention to prevent progression of liver disease and associated comorbidities.

Few pediatric studies have examined the diagnostic overlap between NAFLD and MASLD, and findings have varied depending on methodology and population. Sanchez-Torres et al., using data from the TARGET-NASH network, reported 99% concordance between NAFLD and MASLD based on clinical or biopsy-supported diagnoses; however, in the smaller biopsy subset (n=41), concordance dropped to 90%, suggesting greater diagnostic variability when applying stricter histologic criteria.²³ Using data from the National Health and Nutrition Examination Survey (NHANES), Ma et al. estimated that 85% of U.S. adolescents aged 12–19 years SLD met MASLD criteria when steatosis was defined using controlled attenuation parameter (CAP).²⁴ However, CAP’s moderate accuracy in children may introduce uncertainty. More recently, Noon et al. conducted a larger analysis of adolescents using NHANES data and found that 77.2% of those with elevated ALT met MASLD criteria, while 20.2% had cryptogenic ALT elevation and 2.6% had secondary causes, including viral hepatitis and hepatotoxic medication use.²⁵ These studies highlight how estimates of NAFLD–MASLD concordance in children depend on how steatosis is defined (histology, CAP, or ALT), how metabolic dysfunction is assessed, and whether secondary causes are systematically evaluated. Our findings build on this literature by applying the MASLD framework in a large, rigorously phenotyped cohort with biopsy-confirmed steatosis, centralized histologic review, and systematic assessment of medication exposure and alcohol use.

In adult populations, MASLD–NAFLD concordance is highest in clinic-based cohorts with biopsy-confirmed NAFLD. A U.S. registry study of adults with biopsy-proven NAFLD and rigorous exclusion of secondary causes reported 99.9% concordance with MASLD, while a similar Turkish study reported 99.7%.^26,27 In contrast, population-based estimates are lower. In NHANES III, 94.7% of U.S. adults with NAFLD met MASLD criteria, while two separate analyses of NHANES 2017–2020 data found that 89% of adults with SLD qualified as MASLD.^28,29 Notably, concordance drops further in populations with lower average BMI. Among Indian adults with NAFLD and BMI <23 kg/m², only 83.5% met MASLD criteria, and a northern Chinese study found that 76.9% of biopsy-confirmed NAFLD cases met MASLD criteria, with a mean BMI of 27 kg/m² in the MASLD group.^12,20 These findings reinforce that NAFLD–MASLD concordance depends not only on study design (clinical vs. population-based) and diagnostic methodology (biopsy vs. imaging), but also on population characteristics, especially the burden of metabolic dysfunction. Collectively, these studies suggest that while MASLD provides a robust framework for classifying most cases of NAFLD in adults and children, a nontrivial proportion may be reclassified when comprehensive diagnostic evaluation is performed.

In our pediatric cohort, this same principle holds: although the majority met MASLD criteria, approximately 16% were reclassified into other categories. Notably, 3% of the full cohort and over 5% of adolescents met criteria for Met-ALD. These findings are consistent with national estimates of adolescent drinking. According to the 2023 National Survey on Drug Use and Health, 6.9% of individuals aged 12–17 reported alcohol use in the past month, with 3.9% reporting binge drinking.³⁰ Adolescent alcohol consumption also tends to increase with age. In the 2023 Monitoring the Future report, 13.7% of 10th graders had consumed alcohol in the past 30 days, and 5.4% reported binge drinking; among 12th graders, these rates rose to 24.3% and 10.2%, respectively.³¹ Our reclassification of Met-ALD was conservative, consistent with the AASLD Pediatric MASLD Clinical Practice Guideline, which emphasizes that no amount of alcohol use should be considered safe in adolescents.³² This approach also aligns with the Rinella et al. Delphi consensus, in which 95% of pediatric hepatologists agreed that adult alcohol thresholds are not appropriate for youth.⁶ Taken together, these data support systematic alcohol screening beginning in early adolescence. Importantly, emerging studies suggest that even low levels of alcohol use in adolescents with MASLD increase the risk of progression to cirrhosis and early mortality.^33–35 A near-fatal case of acute liver failure in an adolescent with MASLD and concurrent alcohol use, as described by Schenker et al., illustrates the potential severity of this combination.^36–38 In our study, PNPLA3 genotyping further supported biological differentiation: the high-risk GG genotype was significantly less common in the Met-ALD group (31%) than among those with MASLD (56%), suggesting that alcohol, not underlying genetic susceptibility, was the more salient factor in steatosis development for these adolescents. Current American Academy of Pediatrics guidelines recommend routine alcohol screening beginning at age 12, using validated tools such as the CRAFFT (Car, Relax, Alone, Forget, Friends, Trouble) questionnaire or the AUDIT.¹⁵ These screening recommendations are echoed in the recent AASLD pediatric MASLD clinical practice statement.³² Our findings highlight the urgent need for systematic alcohol screening when evaluating adolescents with MASLD.

DISLD was the most common non-MASLD classification in our cohort, accounting for 6.5% of participants. These children were reclassified based on documented exposure to medications with known or suspected steatogenic effects, including corticosteroids, valproate, methotrexate, and second-generation antipsychotics such as risperidone and olanzapine. Mechanistically, DISLD results from impaired mitochondrial β-oxidation, increased de novo lipogenesis, insulin resistance, and direct hepatocellular toxicity.^39–41 While these medications were not considered exclusionary at the time of NAFLD diagnosis, under the MASLD framework, the presence of a known steatogenic exposure warrants reclassification. This transition reflects a shift in diagnostic philosophy from a diagnosis of exclusion (NAFLD) to a framework built on attribution. Crucially, this reclassification is not a critique of prior clinical decisions. Pediatric gastroenterologists enrolling patients used best available standards and routinely discussed medication-related risk with families. Our retrospective reclassification relied on a structured framework applied to rigorous histologic and exposure data, enabling attribution to the most likely associated factor. The biological distinction of DISLD is further supported by genotype data: the PNPLA3 GG genotype was present in only 22% of children with DISLD, as compared to 56% in those classified as MASLD. Thus, these are not simply overlapping phenotypes but may represent distinct mechanistic pathways. As the use of steatogenic medications continues to rise in pediatric mental health and autoimmune care, awareness of DISLD will become increasingly critical in both diagnosis and management of pediatric liver disease.^42,43

Finally, a small but clinically important subgroup of children were found to have <5% macrovesicular steatosis on centralized histologic review, despite having been diagnosed with NAFLD at enrollment based on local pathology and clinical context. These cases were reclassified as not having SLD under the MASLD framework. The ≥5% macrovesicular fat threshold remains the field standard for both NAFLD and MASLD, reflecting the upper limit of normal hepatic fat content and the level at which steatosis is believed to confer hepatic and metabolic risk.^44,45 Nonetheless, these cases highlight how diagnoses can shift based on where a patient falls in relation to a rigid threshold similar to the reclassification seen when blood pressure or HbA1c values cross clinical cutoffs. Including these patients in our analysis helps illustrate how strict application of research definitions may diverge from clinical practice, particularly at the margins. These findings highlight the need for definitions that are both rigorous and adaptable to clinical context, especially as MASLD terminology is adopted across research and care settings.

Strengths and Limitations

This study was strengthened by its systematic application of the SLD framework to a large, well-characterized cohort of children with biopsy-proven NAFLD. Centralized histologic review by expert pathologists minimized interobserver variability and ensured consistent application of diagnostic thresholds. The inclusion of standardized medical histories, detailed medication data, and validated alcohol use assessments enabled accurate identification of DISLD and Met-ALD, two underrecognized contributors to pediatric steatotic liver disease. Importantly, this analysis included participants from 15 U.S. pediatric liver centers, yielding a demographically diverse cohort representative of real-world clinical settings. Site-level differences accounted for only a modest proportion of the variance in classification outcomes, further supporting the generalizability of findings across centers.

Despite these strengths, several limitations merit consideration. While liver biopsy remains the diagnostic reference standard, its use limits generalizability to settings where noninvasive tools such as imaging or biomarkers are more commonly employed. Alcohol use was assessed with validated instruments and quantified to differentiate MASLD from Met-ALD and ALD, but reliance on self-report may underestimate true exposure. Medication data were collected systematically, and DISLD classification required documented exposure to a steatogenic agent prior to biopsy; however, causality could not be confirmed given the retrospective design, and some exposures may have been missed. Finally, although PNPLA3 genotype data provided insight into biological differences across subgroups, other genetic and environmental contributors to disease heterogeneity were not assessed. Future studies using prospective designs and multimodal diagnostics will be needed to further evaluate the MASLD framework in diverse pediatric populations.

Conclusion

In this large, multicenter cohort of children with biopsy-confirmed NAFLD, 84% met criteria for MASLD, while 16% were reclassified as Met-ALD, DISLD, or cryptogenic SLD. Children classified as MASLD exhibited a high burden of cardiometabolic risk, with nearly three qualifying risk factors on average. Approximately 1 in 20 adolescents were reclassified as Met-ALD, reinforcing the importance of routine alcohol-use screening using validated tools. In addition, 6.5% of participants were reclassified as DISLD due to exposure to medications such as antipsychotics, corticosteroids, and antiepileptics, highlighting the growing impact of neuropsychiatric and autoimmune comorbidities in pediatric liver care. Collectively, these findings support the need for comprehensive diagnostic protocols that evaluate both metabolic and non-metabolic contributors to steatosis. As the MASLD framework is increasingly adopted, its careful application in pediatric populations will be essential for ensuring accurate diagnosis, appropriate risk stratification, and the design of inclusive clinical trials.

Table 2.

Characteristics of adolescents age 12 to 17 classified as metabolic dysfunction-associated steatotic liver disease or metabolic dysfunction and alcohol associated steatotic liver disease

Participant characteristics	MASLD (n=494)	Met-ALD (N=33)	P-value
Age, yr, mean (SD)	14.7 (1.8)	16.2 (1.5)	<0.001
Sex, N (%)			0.73
Male	358 (72)	23 (70)
Female	136 (28)	10 (30)
Ethnicity, N (%)			0.02
Hispanic	348 (70)	17 (52)
Anthropometrics
Weight, kg, mean (SD)	94.2 (22.7)	109.2 (27.6)	<0.001
Height, cm, mean (SD)	1.7 (0.1)	1.7 (0.1)	0.009
BMI, mean (SD)	33.9 (6.3)	37.4 (8.0)	0.002
BMI Z-score, mean (SD)	2.8 (1.2)	3.3 (1.7)	0.03
Systolic blood pressure, mm Hg, mean (SD)	123 (13)	124 (10)	0.70
Diastolic blood pressure, mm Hg, mean (SD)	69 (9)	68 (9)	0.33
Laboratory Values
Aspartate aminotransferase, U/L, mean (SD)	61 (48)	62 (46)	0.66
Alanine aminotransferase, U/L, mean (SD)	108 (94)	104 (75)	0.73
Alkaline phosphatase, U/L, mean (SD)	185 (96)	132 (81)	<0.001
γ-glutamyl transpeptidase, U/L, mean (SD)	45 (36)	64 (106)	0.02
Total bilirubin, mg/dL, mean (SD)	0.5 (0.3)	0.6 (0.4)	0.35
Platelet count, 103 cells/mcl, mean (SD)	290 (75)	267 (58)	0.08
Triglycerides, mg/dL, mean (SD)	153 (79)	157 (103)	0.80
Total cholesterol, mg/dL, mean (SD)	165 (37)	156 (39)	0.19
HDL, mg/dL, mean (SD)	38 (8)	37 (12)	0.47
LDL, mg/dL, mean (SD)	98 (32)	91 (32)	0.28
Glucose, mg/dL, mean (SD)	94 (27)	90 (9)	0.43
HbA1c, %, mean (SD)	5.6 (1.0)	5.5 (0.5)	0.64
PNPLA3 GG genotype, N (%)	166 (49%)	9 (31%)	0.03
Histology
Definite Steatohepatitis, N (%)	117 (24%)	9 (27%)	0.64
Fibrosis stages 2–4	133 (27%)	8 (24%)	0.72
Ballooning, Grade 2 (many), N (%)	55 (11%)	7 (21%)	0.08

This table summarizes demographic, anthropometric, laboratory, genetic, and histologic characteristics of adolescents reclassified as having MASLD or Met-ALD. Comparisons are restricted to participants aged ≥12 years with available alcohol intake data. Continuous variables were compared using independent samples t-tests; categorical variables were compared using Pearson’s chi-square tests. Due to non-normal distribution, liver enzyme values were compared using Kruskal-Wallis tests.

List of Abbreviations:

ALT

Alanine aminotransferase

AUDIT

Alcohol Use Disorders Identification Test

AUDIT-C

Alcohol Use Disorders Identification Test - Consumption

ALD

Alcohol-associated liver disease

AST

Aspartate aminotransferase

BMI

Body mass index

CMRF

Cardiometabolic risk factor

CDC

Center for Disease Control

CRAFFT

Car, Relax, Alone, Forget, Friends, Trouble

DISLD

Drug-induced steatotic liver disease

GGT

Gamma-glutamyl transferase

HbA1c

Hemoglobin A1c

HDL

High-density lipoprotein

LDL

Low-density lipoprotein

Met-ALD

Metabolic dysfunction and alcohol associated steatotic liver disease

MASLD

Metabolic dysfunction-associated steatotic liver disease

NHANES

National Health and Nutrition Examination Survey

NAFLD

Nonalcoholic fatty liver disease

NASH CRN

Nonalcoholic Steatohepatitis Clinical Research Network

PNPLA3

Patatin-like phospholipase domain containing 3

SLD

Steatotic liver disease