Abstract
Objectives:
To characterize longitudinal patterns of alcohol use and evaluate associations with liver enzyme changes in youth with metabolic dysfunction–associated steatotic liver disease (MASLD).
Methods:
We analyzed 1172 children with biopsy-confirmed MASLD and longitudinal Alcohol Use Disorders Identification Test–Consumption (AUDIT-C) data. Group-based trajectory modeling identified alcohol use patterns over time. Linear mixed models assessed associations between alcohol trajectory groups and alanine aminotransferase (ALT), aspartate aminotransferase (AST), and gamma-glutamyl transferase (GGT). Generalized linear mixed models evaluated predictors of alcohol use, including age, sex, body mass index (BMI), fibrosis stage, and household income.
Results:
Among participants diagnosed with MASLD in childhood, 206 (17.6%) reported alcohol use during adolescence or early adulthood. Three alcohol-use groups were identified: non-users, adolescent initiators, and adult initiators. Heavy episodic drinking was reported by 48 of 71 adolescent initiators (67.6%) and 53 of 135 adult initiators (39.3%). Compared with non-users, alcohol initiators had lower liver chemistry values at enrollment but greater annual increases during follow-up. Adolescent initiators had a 4% greater annual increase in ALT, while adult initiators had 6%, 4%, and 3% greater annual increases in ALT, AST, and GGT, respectively.
Conclusions:
Nearly 1 in 5 youth with MASLD reported alcohol use during adolescence or early adulthood despite ongoing subspecialty care. Alcohol-use trajectories were associated with greater annual increases in liver chemistries, particularly among participants with heavy episodic drinking. These findings support routine alcohol screening and developmentally tailored anticipatory guidance in pediatric MASLD care.
Keywords: adolescent behavior, AUDIT-C, biomarkers, hepatocellular injury, longitudinal studies, pediatric liver disease, risk factors
INTRODUCTION
Metabolic dysfunction–associated steatotic liver disease (MASLD) is the most common chronic liver disease in children and adolescents in the United States.1,2 Although MASLD is typically diagnosed in the absence of significant alcohol consumption, the 2025 American Association for the Study of Liver Diseases (AASLD) Clinical Practice Statement emphasizes that alcohol use is an important and underrecognized modifier of liver injury in affected youth. 3 Adolescence is a developmental period marked by rapid increases in experimentation and risk behaviors, including alcohol use.4,5 National surveillance data indicate that ~60% of United States adolescents have consumed alcohol, and the prevalence of regular drinking rises sharply between ages 13 and 18. 4 For youth with underlying liver disease, this pattern is concerning given alcohol’s well-established hepatotoxicity and its potential to worsen underlying liver injury. 6
In adults with MASLD, even low or moderate levels of alcohol exposure are associated with faster fibrosis progression, increased risk of hepatocellular carcinoma, and elevated all-cause mortality. 6 Whether similar susceptibility to alcohol-associated liver injury exists earlier in life remains unclear. In particular, the timing of alcohol initiation, the developmental trajectories of use, and the extent to which alcohol may influence liver health in adolescents with MASLD remain poorly characterized. These gaps are particularly salient in the context of rising alcohol-associated liver disease (ALD) among adolescents and young adults, with the steepest increases observed among females. 7 Despite these trends, few studies have examined longitudinal patterns of alcohol use or their potential hepatologic consequences in youth with MASLD.
To address this gap, we analyzed data from a large, multicenter cohort of adolescents with biopsy-confirmed MASLD to (1) delineate trajectories of alcohol use from adolescence into early adulthood and (2) evaluate associations between alcohol-use trajectories and longitudinal changes in liver chemistries. We hypothesized that alcohol use would increase with age and that alcohol-use trajectories characterized by earlier initiation or greater exposure would be associated with greater longitudinal increases in alanine aminotransferase (ALT), aspartate aminotransferase (AST), and gamma-glutamyl transferase (GGT).
METHODS
Study design and population
We conducted a longitudinal cohort analysis of children and adolescents with metabolic dysfunction–associated steatotic liver disease (MASLD) enrolled in the Nonalcoholic Steatohepatitis Clinical Research Network (NASH CRN) between 2004 and 2020. Participants were recruited from 15 pediatric centers across the United States. Eligibility criteria included age <18 years at enrollment, biopsy-confirmed steatotic liver disease, and the presence of at least one cardiometabolic risk factor consistent with the MASLD framework (described below) (Figure 1). Participants without at least one follow-up visit were excluded. All participants met criteria for MASLD at baseline enrollment, and analyses focused on longitudinal patterns of alcohol use and subsequent changes in liver chemistries following diagnosis.
FIGURE 1.
Flow diagram for cohort selection. From 1479 children in the NASH CRN, we excluded participants who did not meet criteria for MASLD, were younger than 8 years at enrollment, or lacked a follow-up visit. The remaining 1172 participants formed the analytic cohort. Abbreviations: MASLD, metabolic dysfunction–associated steatotic liver disease; NASH CRN, Nonalcoholic Steatohepatitis Clinical Research Network.
Ethical approval for this study was provided by the Institutional Review Board at each participating NASH CRN site. All research procedures were conducted in accordance with the ethical principles outlined in the Declaration of Helsinki and the Declaration of Istanbul. Written informed consent was obtained from a parent or legal guardian for all participants, and written assent was obtained from children aged 8 years and older.
MASLD diagnosis
MASLD was defined as the presence of hepatic steatosis accompanied by at least 1 of 5 cardiometabolic risk factors. 3 These included: (1) body mass index (BMI) at or above the 85th percentile for age and sex (BMI z-score ≥ +1); (2) type 2 diabetes or pre-diabetes, including fasting serum glucose levels ≥100 mg/dL, hemoglobin A1c (HbA1c) ≥5.7%, or a diagnosis or treatment of type 2 diabetes; (3) blood pressure at or above the 95th percentile for children younger than 13 years of age, or ≥130/85 mm Hg for those aged 13 years or older, or the use of antihypertensive medication; (4) elevated plasma triglycerides, defined as ≥100 mg/dL for children younger than 10 years of age or ≥150 mg/dL for those aged 10 years or older, or treatment with lipid-lowering therapy; and (5) high-density lipoprotein cholesterol levels ≤40 mg/dL or treatment with lipid-lowering therapy.
Exposure assessment: Alcohol use
As part of standard clinical care for MASLD, alcohol use was routinely assessed, and counseling regarding alcohol avoidance was provided at clinical visits. 3 Alcohol use was assessed using the three-item Alcohol Use Disorders Identification Test–Consumption (AUDIT-C) questionnaire, administered at each research visit for participants aged ≥12 years. 8 The AUDIT-C evaluates (1) drinking frequency, (2) typical drinking quantity, and (3) frequency of heavy episodic drinking. Response options range from “never” to “four or more times per week” for drinking frequency, from “1–2 drinks” to “10 or more drinks” for typical quantity, and from “never” to “daily or almost daily” for consumption of 6 or more drinks on one occasion. Each item is scored from 0 to 4, yielding a total score ranging from 0 to 12. Heavy episodic drinking (HED) was defined using AUDIT-C item 3; any response greater than “never” was classified as HED for that visit. For participant-level summaries, ever HED was defined as any report of HED at one or more study visits during follow-up. Interim alcohol-use history was also collected at each study visit. Objective biomarkers of alcohol use, including blood phosphatidylethanol, were not measured as part of this study.
Outcome measures: Liver chemistries
The primary outcomes were repeated measures of ALT, AST, and GGT, obtained from fasting blood samples collected at enrollment and annually during follow-up. These liver chemistries were selected because they were collected uniformly across NASH CRN sites and represent standard clinical markers used for longitudinal monitoring in pediatric MASLD. The AASLD Pediatric MASLD Clinical Practice Statement recommends ALT, AST, and GGT for follow-up in children with MASLD. 3 In prior pediatric studies, dynamic changes in ALT and GGT were associated with histologic change, including both improvement and worsening, and these findings were reproduced across multiple trials with paired biopsy data.9,10 AST was included because it provides complementary information about hepatocellular injury and has been associated with fibrosis severity.11,12 In adults, AST is also associated with alcohol use. 13
Measurement of covariates
Demographic characteristics were collected at enrollment using structured interviews and included age, sex, race, and ethnicity. Sex was categorized as male or female. Race was self-identified as American Indian or Alaska Native, Asian, Black, Native Hawaiian or Pacific Islander, White, Multiracial, or Unknown, and ethnicity was categorized as Hispanic or non-Hispanic.
Anthropometric measurements were obtained at baseline and each annual visit using standardized NASH CRN protocols. Height and weight were measured without shoes, and BMI was calculated as kilograms per meter squared. BMI z-scores were derived using the CDC 2000 LMS method with age and sex-specific reference values. 14
Laboratory covariates included fasting glucose, insulin, hemoglobin A1c, hemoglobin, platelet count, total cholesterol, HDL cholesterol, LDL cholesterol, and triglycerides. Liver histology was assessed at enrollment according to NASH CRN criteria and was used to characterize baseline disease severity at diagnosis. Repeat liver biopsy was not performed systematically during follow-up. Histologic features included steatosis grade, lobular inflammation, portal inflammation, and hepatocellular ballooning. Fibrosis stage was categorized as stage 0, stage 1a, stage 1b, stage 1c, stage 2, stage 3, or stage 4 based on the NASH CRN scoring system.15,16
Statistical analysis
All analyses were conducted using SAS Studio version 9.4 (SAS Institute Inc., Cary, North Carolina). Statistical tests with two-sided and a p-value of <0.05 were considered statistically significant.
Descriptive analyses
Baseline characteristics were summarized overall and stratified by alcohol use groups, which included 2 trajectory-defined alcohol use patterns and a non-use group. Continuous variables were summarized as means with standard deviations, and categorical variables were summarized as frequencies and percentages. Differences between groups were assessed using analysis of variance or Kruskal–Wallis for continuous variables and chi-square tests for categorical variables.
Alcohol trajectory modeling
To identify longitudinal patterns of alcohol use between ages 12 and 24 years, we applied group-based trajectory modeling to AUDIT-C scores using the PROC TRAJ procedure. 17 This method identifies subgroups of participants who follow similar developmental patterns of alcohol use over time and is well-suited for describing heterogeneous behavioral trajectories in adolescents. 17 Participants who never reported alcohol use (AUDIT-C score = 0 at all visits) were prespecified as a distinct non-use group and were not included in trajectory estimation. Trajectory modeling was therefore applied only among participants who reported any alcohol use (AUDIT-C score ≥1 at least once), and the resulting trajectory groups were subsequently combined with the non-use group for descriptive and inferential analyses.
Age at each visit was rounded to the nearest quarter year to improve temporal resolution. When more than one AUDIT-C record occurred at the same age, the earliest complete record was retained. Candidate models assumed a censored normal distribution and included 2 to 4 trajectory groups, with polynomial terms allowed to vary by group to capture nonlinear patterns. Model selection among ever-users was guided by Bayesian information criterion (BIC) values and clinical interpretability. 18
Associations between alcohol-use trajectories and liver chemistries
Linear mixed models were used to estimate associations between alcohol-use trajectory group and longitudinal ALT, AST, and GGT values. Liver chemistry values were log-transformed to improve model fit and approximate normality. Time was modeled as age at visit, allowing liver chemistry change to be evaluated across developmental age. Models included alcohol-use trajectory group, age at visit, and an interaction between alcohol-use trajectory group and age at visit to evaluate whether changes in liver chemistries over time differed by alcohol-use group. Because alcohol-use trajectories were also defined across developmental age, modeling liver chemistry values as a function of age at visit aligned the outcome models with the developmental time scale used to define alcohol-use patterns.
Participant-specific random intercepts and slopes for age at visit were included to account for within-person correlation and individual variability in enrollment liver chemistry values and change over time. Clinic random effects were explored but could not be reliably estimated in models that also included participant-specific random intercepts and slopes; therefore, clinic effects were not included in the final models. Variance components, compound symmetry, unstructured, and autoregressive covariance structures were compared using Akaike Information Criterion values, and the structure with the lowest value was selected for each model.
Models adjusted for enrollment age, sex, race, ethnicity, baseline fibrosis stage, baseline lobular inflammation, and time-varying BMI. Results are presented as exponentiated estimates. The alcohol-use group main effect is interpreted as the adjusted ratio of liver chemistry values at enrollment compared with non-users. The alcohol-use group-by-time interaction is interpreted as the relative annual change in liver chemistry values compared with non-users. Linear mixed models incorporate all available outcome data under a missing-at-random assumption using maximum likelihood estimation, allowing participants with incomplete follow-up to remain in the analysis.
Determinants of alcohol use over time
In a complementary exploratory analysis to identify factors associated with alcohol use across adolescence, we fitted generalized linear mixed models with a logit link and any alcohol use at each visit (yes or no) as the outcome. Any alcohol use at a given visit was classified as yes if the AUDIT-C score was ≥1 and no if the score was 0. This approach supports our objective of identifying the onset of alcohol use and reinforces current clinical guidance that MASLD cannot be diagnosed in adolescents in the presence of any alcohol exposure. 3 Random intercepts and random slopes were included to account for individual differences in baseline probability of use and rate of change in use over time. An unstructured covariance matrix was selected based on Akaike Information Criterion values. Covariates included time-varying age and BMI, as well as baseline sex, race and ethnicity, fibrosis stage, and household income.
Sensitivity analysis
Using the same linear mixed model framework, we conducted a sensitivity analysis in which alcohol initiators were further stratified by ever HED status. The resulting categories were non-users, adolescent initiators without HED, adolescent initiators with ever HED, adult initiators without HED, and adult initiators with ever HED. This analysis evaluated whether associations between alcohol-use trajectories and longitudinal liver chemistry changes differed according to heavy episodic drinking.
Because glycemic control may confound associations between alcohol-use trajectories and liver chemistry changes, we repeated the primary models with additional adjustment for time-varying HbA1c. 19 We also examined models including baseline diabetes status, defined using available diabetes diagnosis, diabetes medication use, HbA1c, and fasting glucose data, to assess whether diabetes status materially changed the alcohol-use trajectory estimates.
RESULTS
Study population
Among 1479 children enrolled in the NASH CRN, 1172 met the inclusion criteria for this analysis. Participants were excluded for the absence of steatotic liver disease, failure to meet MASLD criteria, or lack of follow-up (Figure 1). The cohort had a mean age at enrollment of 13.0 years (SD 2.7) and a median follow-up duration of 3.1 years (IQR 1.8, 5.4). Most participants were male and identified as Hispanic. At enrollment, liver enzymes were elevated, with a median ALT of 83.0 U/L, AST of 49.0 U/L, and GGT of 36.0 U/L. Definite steatohepatitis was present in 23.0% of the cohort. Fibrosis was present in 68.0% of participants at baseline, including bridging fibrosis in 12.0% and cirrhosis in 1.4%. 1
Alcohol use trajectories
Group-based trajectory modeling was applied among participants who reported any alcohol use (AUDIT-C score ≥1 at least once), while participants who never reported alcohol use were prespecified as a distinct non-use group. Among ever-users, models with 2 through 4 trajectory groups were evaluated. A two-group linear model with first-order terms provided the best statistical fit (BIC −1486.90). However, when combined with the prespecified non-use group, a three-group structure consisting of non-users and 2 alcohol initiation trajectories provided the most clinically interpretable solution and was retained for subsequent analyses (Figure 2).
FIGURE 2.
Longitudinal alcohol trajectories in youth with MASLD (ages 12–24 y). Observed and model-predicted AUDIT-C scores are shown by alcohol-use trajectory group: non-users (teal; n = 966), adolescent initiators (orange; n = 71), and adult initiators (purple; n = 135). Points represent observed mean AUDIT-C scores at each age, solid lines represent group-based trajectory model–predicted AUDIT-C scores, and shaded ribbons represent 95% confidence intervals for the predicted trajectories. The vertical dashed line at age 21 denotes the United States legal drinking age. AUDIT-C scores range from 0 to 12, with higher scores indicating greater alcohol use. MASLD, metabolic dysfunction–associated steatotic liver disease; AUDIT-C, Alcohol Use Disorders Identification Test–Consumption.
The non-use group included 966 participants (82.4%) who did not report alcohol use during follow-up. The adolescent initiation trajectory included 71 participants (6.1%) who first reported alcohol use at ~14 years of age and later demonstrated the steepest rise in AUDIT-C scores, reaching a mean of 6.1 by age 21 years or older. The adult initiation trajectory included 135 participants (11.5%) who first reported use at ~18 years of age and displayed a modest but consistent increase in AUDIT-C scores thereafter (Figure 2).
Participant characteristics differed by alcohol-use group. At enrollment, non-users were younger than adolescent and adult initiators, with mean ages of 12.7, 14.4, and 14.4 years, respectively (p<0.001). Non-users also had shorter follow-up, with a median follow-up of 2.7 years compared with 4.5 years among adolescent initiators and 6.4 years among adult initiators (p<0.001). Enrollment liver chemistries were higher among non-users, including median AST of 51.0 U/L compared with 42.5 U/L and 44.0 U/L among adolescent and adult initiators, respectively (p<0.001), and median ALT of 85.0 U/L compared with 74.5 U/L and 74.0 U/L, respectively (p=0.006). Advanced fibrosis was also more common among non-users, occurring in 14.6% compared with 8.4% of adolescent initiators and 7.5% of adult initiators (p=0.01). Heavy episodic drinking was reported at least once by 101 participants who initiated alcohol use, including 48 of 71 adolescent initiators (67.6%) and 53 of 135 adult initiators (39.3%). Full comparisons across groups are provided in Table 1.
TABLE 1.
Participant characteristics of 1172 children with metabolic dysfunction–associated steatotic liver disease, according to alcohol-use group
| Alcohol-use groups | |||||
|---|---|---|---|---|---|
| Non-users (n=966) | Adolescent initiators (n=71) | Adult initiators (n=135) | Total (N=1172) | p | |
| Age (years) | <0.001 a | ||||
| Mean (SD) | 12.7 (2.7) | 14.4 (2.3) | 14.4 (2.5) | 13.0 (2.7) | |
| Follow-up (years) | <0.001 b | ||||
| Median (IQR) | 2.7 (1.6, 4.7) | 4.5 (2.3, 7.0) | 6.4 (4.1, 10.5) | 3.1 (1.8, 5.4) | |
| HED, n (%) | <0.001 c | ||||
| Ever HED | 0 (0.0%) | 48 (67.6%) | 53 (39.3%) | 101 (8.6%) | |
| No HED | 966 (100.0%) | 23 (32.4%) | 82 (60.7%) | 1071 (91.4%) | |
| Sex, n (%) | 0.54 c | ||||
| Female | 269 (27.8%) | 18 (25.4%) | 43 (31.9%) | 330 (28.2%) | |
| Male | 697 (72.2%) | 53 (74.6%) | 92 (68.1%) | 842 (71.8%) | |
| Race and ethnicity, n (%) | 0.80 c | ||||
| Hispanic | 694 (71.8%) | 49 (69.0%) | 98 (72.6%) | 841 (71.8%) | |
| Non-Hispanic White | 213 (22.0%) | 19 (26.8%) | 31 (23.0%) | 263 (22.4%) | |
| Non-Hispanic Other | 59 (6.1%) | 3 (4.2%) | 6 (4.4%) | 68 (5.8%) | |
| BMI z-score | 0.64 a | ||||
| Mean (SD) | 2.4 (0.8) | 2.4 (0.8) | 2.4 (0.8) | 2.4 (0.8) | |
| Systolic blood pressure (mm Hg) | 0.55 b | ||||
| Median (IQR) | 88.0 (68.0, 97.0) | 88.0 (55.0, 96.0) | 84.5 (63.0, 97.0) | 88.0 (68.0, 97.0) | |
| Diastolic blood pressure (mm Hg) | 0.004 b | ||||
| Median (IQR) | 66.0 (44.0, 85.0) | 54.5 (34.0, 70.0) | 63.5 (37.0, 82.0) | 65.0 (42.0, 84.0) | |
| GGT (U/L) | 0.23 b | ||||
| Median (IQR) | 36.0 (25.0, 55.0) | 33.0 (23.0, 50.0) | 33.5 (24.0, 52.0) | 36.0 (25.0, 54.0) | |
| AST (U/L) | <0.001 b | ||||
| Median (IQR) | 51.0 (37.5, 74.0) | 42.5 (32.0, 53.0) | 44.0 (31.0, 65.0) | 49.0 (36.0, 71.0) | |
| ALT (U/L) | 0.006 b | ||||
| Median (IQR) | 85.0 (57.0, 139.0) | 74.5 (49.0, 108.0) | 74.0 (46.0, 110.0) | 83.0 (55.0, 134.0) | |
| Total cholesterol (mg/dL) | 0.74 a | ||||
| Mean (SD) | 165.5 (37.6) | 162.5 (40.1) | 166.8 (37.9) | 165.4 (37.7) | |
| HDL cholesterol (mg/dL) | 0.86 b | ||||
| Median (IQR) | 38.0 (33.0, 44.0) | 38.5 (33.0, 47.0) | 38.0 (33.0, 44.0) | 38.0 (33.0, 44.0) | |
| LDL cholesterol (mg/dL) | 0.52 a | ||||
| Mean (SD) | 97.3 (30.8) | 94.2 (28.7) | 99.4 (31.2) | 97.4 (30.7) | |
| Triglycerides (mg/dL) | 0.18 b | ||||
| Median (IQR) | 134.0 (94.0, 187.0) | 113.5 (79.0, 173.0) | 129.0 (89.0, 175.0) | 133.0 (92.0, 184.0) | |
| HbA1c (%) | 0.31 b | ||||
| Median (IQR) | 5.4 (5.1, 5.6) | 5.4 (5.2, 5.6) | 5.4 (5.2, 5.7) | 5.4 (5.1, 5.6) | |
| Fasting serum glucose (mg/dL) | <0.001 b | ||||
| Median (IQR) | 89.0 (83.0, 96.0) | 87.0 (80.0, 95.0) | 84.0 (79.0, 91.0) | 88.0 (82.0, 95.0) | |
| Insulin (mU/mL) | 0.67 b | ||||
| Median (IQR) | 25.5 (14.9, 43.5) | 25.5 (16.0, 43.5) | 29.0 (18.0, 41.0) | 25.9 (15.0, 42.8) | |
| HOMA-IR | 0.98 b | ||||
| Median (IQR) | 99.0 (57.2, 175.5) | 99.6 (53.9, 165.8) | 106.1 (64.8, 154.9) | 100.9 (58.7, 168.5) | |
| Hemoglobin (g/dL) | 0.01 a | ||||
| Mean (SD) | 13.8 (1.1) | 14.2 (1.6) | 14.1 (1.3) | 13.9 (1.2) | |
| Platelets (×103 cells/µL) | 0.67 b | ||||
| Median (IQR) | 287.0 (253.0, 333.0) | 298.0 (252.5, 335.5) | 286.0 (251.0, 328.0) | 287.0 (253.0, 333.0) | |
| Steatosis amount, n (%) | 0.09 c | ||||
| 5%–33% | 248 (25.7%) | 22 (31.0%) | 35 (25.9%) | 305 (26.0%) | |
| >33%–66% | 280 (29.0%) | 27 (38.0%) | 48 (35.6%) | 355 (30.3%) | |
| >66% | 438 (45.3%) | 22 (31.0%) | 52 (38.5%) | 512 (43.7%) | |
| Lobular inflammation—number of foci under ×20 magnification, n (%) | 0.27 c | ||||
| <2 | 537 (55.6%) | 42 (59.2%) | 81 (60.0%) | 660 (56.3%) | |
| 2–4 | 373 (38.6%) | 23 (32.4%) | 42 (31.1%) | 438 (37.4%) | |
| >4 | 56 (5.8%) | 6 (8.5%) | 12 (8.9%) | 74 (6.3%) | |
| Hepatocellular ballooning, n (%) | 0.33 c | ||||
| None | 609 (63.0%) | 42 (59.2%) | 77 (57.0%) | 728 (62.1%) | |
| Few | 249 (25.8%) | 21 (29.6%) | 35 (25.9%) | 305 (26.0%) | |
| Many | 108 (11.2%) | 8 (11.3%) | 23 (17.0%) | 139 (11.9%) | |
| Portal inflammation, n (%) | 0.74 c | ||||
| None | 105 (10.9%) | 11 (15.5%) | 13 (9.6%) | 129 (11.0%) | |
| Mild | 715 (74.1%) | 51 (71.8%) | 103 (76.3%) | 869 (74.2%) | |
| More than mild | 145 (15.0%) | 9 (12.7%) | 19 (14.1%) | 173 (14.8%) | |
| Fibrosis stage, n (%) | 0.01 c | ||||
| 0: none | 290 (30.1%) | 33 (46.5%) | 51 (38.1%) | 374 (32.0%) | |
| 1a: mild, zone 3 perisinusoidal | 70 (7.3%) | 6 (8.5%) | 10 (7.5%) | 86 (7.4%) | |
| 1b: moderate, zone 3 perisinusoidal | 34 (3.5%) | 4 (5.6%) | 9 (6.7%) | 47 (4.0%) | |
| 1c: portal/periportal only | 293 (30.5%) | 14 (19.7%) | 31 (23.1%) | 338 (29.0%) | |
| 2: zone 3 and periportal, any combination | 135 (14.0%) | 8 (11.3%) | 23 (17.2%) | 166 (14.2%) | |
| 3: bridging | 126 (13.1%) | 4 (5.6%) | 10 (7.5%) | 140 (12.0%) | |
| 4: cirrhosis | 14 (1.5%) | 2 (2.8%) | 0 (0.0%) | 16 (1.4%) | |
| Steatohepatitis diagnosis, n (%) | 0.01 c | ||||
| No | 266 (27.5%) | 30 (42.3%) | 43 (31.9%) | 339 (28.9%) | |
| Borderline zone 1 pattern | 317 (32.8%) | 12 (16.9%) | 35 (25.9%) | 364 (31.1%) | |
| Borderline zone 3 pattern | 171 (17.7%) | 10 (14.1%) | 18 (13.3%) | 199 (17.0%) | |
| Definite | 212 (21.9%) | 19 (26.8%) | 39 (28.9%) | 270 (23.0%) | |
| PNPLA3 genotype, n (%) | 0.10 c | ||||
| CC | 91 (12.3%) | 11 (16.4%) | 26 (20.8%) | 128 (13.8%) | |
| CG | 256 (34.7%) | 23 (34.3%) | 34 (27.2%) | 313 (33.7%) | |
| GG | 391 (53.0%) | 33 (49.3%) | 65 (52.0%) | 489 (52.6%) | |
ANOVA F-test p-value.
Kruskal–Wallis p-value.
Fisher exact p-value.
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; GGT, gamma-Glutamyl transpeptidase; HbA1c, hemoglobin A1C; HED, heavy episodic drinking; HDL, high-density lipoprotein; HOMA-IR, homeostatic model assessment for insulin resistance; LDL, low-density lipoprotein; PNPLA3, patatin-like phospholipase domain-containing protein 3.
Association between alcohol‑use trajectories and change in liver chemistries
In linear mixed models adjusted for enrollment age, sex, race, ethnicity, baseline fibrosis stage, baseline lobular inflammation, and time-varying BMI, alcohol-use trajectories were associated with differences in longitudinal ALT, AST, and GGT patterns.
Compared with non-users, adolescent initiators had lower adjusted ALT and AST values at enrollment, with ALT 49% lower (95% CI, 10%–71% lower) and AST 39% lower (95% CI, 5%–60% lower). Adult initiators also had lower adjusted enrollment liver chemistries, with ALT 64% lower (95% CI, 47%–75% lower), AST 50% lower (95% CI, 33%–62% lower), and GGT 35% lower (95% CI, 16%–50% lower) compared with non-users.
Despite lower adjusted liver chemistry values at enrollment, alcohol initiators had greater relative annual increases in liver chemistries during follow-up. Adolescent initiators had a 4% greater annual increase in ALT compared with non-users (95% CI, 1%–8% greater). Adult initiators had a 6% greater annual increase in ALT (95% CI, 4%–9% greater), a 4% greater annual increase in AST (95% CI, 2%–6% greater), and a 3% greater annual increase in GGT (95% CI, 1%–4% greater) compared with non-users (Table 2). Model-estimated adjusted ratios over time are shown in Figure 3.
TABLE 2.
Baseline differences and annual changes in liver chemistries by alcohol-use group among 1172 children with metabolic dysfunction–associated steatotic liver disease
| Panel/alcohol-use groups | ALT a | AST a | GGT a |
|---|---|---|---|
| Panel A—Group difference at baseline | |||
| Adolescents initiators | 0.51 (0.29, 0.90) | 0.61 (0.40, 0.95) | 0.72 (0.49, 1.07) |
| Adult initiators | 0.36 (0.25, 0.53) | 0.50 (0.38, 0.67) | 0.65 (0.50, 0.84) |
| Non-users | Ref | Ref | Ref |
| Panel B—Difference in annual change | |||
| Adolescents initiators | 1.04 (1.01, 1.08) | 1.03 (1.00, 1.06) | 1.02 (1.00, 1.04) |
| Adult initiators | 1.06 (1.04, 1.09) | 1.04 (1.02, 1.06) | 1.03 (1.01, 1.04) |
| Non-users | Ref | Ref | Ref |
Boldface indicates statistical significance, defined as a 95% confidence interval that does not include 1.00.
Model adjusted for age at visit, sex, race and ethnicity, fibrosis grade at baseline, lobular inflammation at baseline, and BMI at visit.
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; GGT, gamma-glutamyl transpeptidase.
FIGURE 3.
Age-specific adjusted ratios of liver chemistries by alcohol-use trajectory group among youth with MASLD from ages 12 to 24 years. Age-specific adjusted ratios for ALT, AST, and GGT are shown for adolescent initiators and adult initiators relative to non-users. Solid lines represent model-estimated adjusted ratios, points represent age-specific estimated ratios, and shaded ribbons represent 95% confidence intervals. The dashed horizontal line at 1.00 indicates no difference compared with non-users. Ratios >1.00 indicate higher liver chemistry values relative to non-users at the same age, whereas ratios <1.00 indicate lower values. Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; MASLD, metabolic dysfunction–associated steatotic liver disease.
Correlates of any alcohol use over follow-up
In an exploratory longitudinal model evaluating factors associated with reporting any alcohol use during follow-up, each additional year of age was associated with higher odds of alcohol use (OR 1.30; 95% CI, 1.21–1.40). Advanced fibrosis at baseline was associated with lower odds of reporting alcohol use (OR 0.31; 95% CI, 0.13–0.76) compared with no fibrosis, whereas mild-to-moderate fibrosis was not associated with alcohol use. Household income greater than $50,000 per year was associated with higher odds of reporting alcohol use (OR 2.78; 95% CI, 1.46–5.29). No clear associations were observed for sex, race, or ethnicity (Table 3).
TABLE 3.
Generalized linear mixed model of characteristics associated with alcohol use among 1172 children with metabolic dysfunction–associated steatotic liver disease
| Model | Odds ratio (95% CI) |
|---|---|
| Age | 1.30 (1.21, 1.40) |
| Sex | |
| Female | 1.18 (0.67, 2.09) |
| Male | Ref |
| Race and ethnicity | |
| Non-Hispanic Other | 0.40 (0.12, 1.35) |
| Non-Hispanic White | 0.65 (0.34, 1.25) |
| Hispanic | Ref |
| Fibrosis stage | |
| Advanced fibrosis | 0.31 (0.13, 0.76) |
| Mild-to-moderate fibrosis | 0.64 (0.36, 1.12) |
| No fibrosis | Ref |
| BMI | 1.00 (0.97, 1.04) |
| Income | |
| $50,000 or higher | 2.78 (1.46, 5.29) |
| Less than $50,000 | Ref |
Boldface indicates statistical significance, defined as a 95% confidence interval that does not include 1.00.
Abbreviations: BMI, body mass index; Ref, reference.
Sensitivity analyses
In the sensitivity analysis stratifying alcohol initiators by HED status, associations between alcohol-use trajectories and annual liver chemistry change were strongest among participants who reported HED. Adolescent initiators without HED did not differ clearly from non-users in adjusted enrollment liver chemistries or annual liver chemistry changes. In contrast, adolescent initiators with ever HED had greater annual increases in ALT, AST, and GGT compared with non-users, including a 5% greater annual increase in ALT (95% CI, 1%–10% greater), a 4% greater annual increase in AST (95% CI, 1%–7% greater), and a 3% greater annual increase in GGT (95% CI, 0%–6% greater) (Supplemental Table S1, https://links.lww.com/HC9/C485).
Among adult initiators, annual increases in liver chemistries were greater in both HED strata, with the largest estimates among those with ever HED. Adult initiators without HED had a 5% greater annual increase in ALT (95% CI, 2%–8% greater) and a 3% greater annual increase in AST (95% CI, 1%–5% greater), while the annual change in GGT was not clearly different from non-users. Adult initiators with ever HED had greater annual increases in all 3 liver chemistries, including ALT (8% greater; 95% CI, 4%–12% greater), AST (6% greater; 95% CI, 3%–8% greater), and GGT (5% greater; 95% CI, 2%–7% greater) (Supplemental Table S1, https://links.lww.com/HC9/C485).
Results were generally similar after additional adjustment for time-varying HbA1c. In these models, the estimate for annual ALT change among adolescent initiators remained positive (4% greater; 95% CI, 0%–7% greater), while annual changes in AST and GGT were not clearly different from non-users. Among adult initiators, annual changes remained greater for ALT (6% greater; 95% CI, 3%–8% greater), AST (4% greater; 95% CI, 2%–6% greater), and GGT (3% greater; 95% CI, 1%–5% greater) compared with non-users (Supplemental Table S2, https://links.lww.com/HC9/C486).
DISCUSSION
In this multicenter longitudinal cohort of children and adolescents with biopsy-confirmed MASLD, we identified 3 patterns of alcohol use across adolescence and early adulthood and evaluated their associations with longitudinal liver chemistry trajectories. Most participants did not report alcohol use during follow-up, but nearly 1 in 5 reported alcohol use, typically during mid-adolescence or early adulthood. Although alcohol initiators had lower liver chemistry values at enrollment than non-users, they demonstrated greater annual increases in liver chemistries over time, with the most consistent differences observed among adult initiators and participants who reported heavy episodic drinking. This baseline pattern likely reflects that alcohol-use groups were defined by subsequent reported alcohol use and that participants with greater apparent disease severity at enrollment were less likely to report alcohol use during follow-up. In HED-stratified analyses, adolescent and adult initiators with ever HED had greater annual increases in liver chemistries compared with non-users, whereas associations were weaker or absent among adolescent initiators without HED. Advanced fibrosis at enrollment was associated with lower odds of reporting alcohol use during follow-up, although the mechanisms underlying this association were not directly measured. Together, these findings characterize alcohol initiation patterns in a high-risk pediatric MASLD population and support an association between alcohol exposure and worsening biochemical markers of liver injury during adolescence and early adulthood.
A meaningful proportion of adolescents with MASLD initiated alcohol use during follow-up, despite their chronic liver condition and ongoing subspecialty care. Although the prevalence of initiation in this cohort was lower than national estimates, nearly 1 in 5 adolescents began drinking during a developmental period when the liver is particularly susceptible to injury. 4 This level of initiation is concerning, given their heightened biological vulnerability and the well-established links between early alcohol use and long-term risks of alcohol use disorder, polysubstance involvement, and adverse psychosocial outcomes.20–22 In the National Comorbidity Survey–Adolescent Supplement, ~60% of United States adolescents aged 13–18 reported lifetime alcohol use, with rates rising from 42.5% at ages 13–14 to more than 78% at ages 17–18. 4 The lower rate observed in our cohort may reflect the impact of regular clinical follow-up, increased disease awareness, or family engagement. However, the fact that initiation still occurred among youth with a serious liver condition suggests that standard counseling practices may not fully address the pressures and contexts that shape adolescent drinking. These findings highlight the need for more proactive, developmentally attuned anticipatory guidance that integrates the biological risks of MASLD with the social realities of adolescence. 23
We hypothesized that alcohol use during adolescence or early adulthood would be associated with worsening liver injury markers in youth with MASLD. In adjusted models, alcohol initiators had lower liver chemistry values at enrollment than non-users but demonstrated greater annual increases during follow-up. Adolescent initiators had a 4% greater annual increase in ALT compared with non-users, while adult initiators had 6%, 4%, and 3% greater annual increases in ALT, AST, and GGT, respectively. These findings suggest that alcohol exposure during adolescence and early adulthood may contribute to worsening hepatocellular injury in a population already susceptible to metabolic liver disease. The association appeared more pronounced among participants who reported heavy episodic drinking. Among adolescent initiators, greater annual increases in ALT, AST, and GGT were observed primarily among those with ever HED, whereas adolescent initiators without HED did not demonstrate clear differences in annual liver chemistry change compared with non-users. Among adult initiators, annual increases were greatest among those with ever HED, although adult initiators without HED also had greater annual increases in ALT and AST. These findings suggest that both the developmental timing of alcohol initiation and drinking patterns may be relevant to liver injury markers in youth with MASLD. These estimates may be conservative given the known underreporting of alcohol use among adolescents and young adults due to age-related alcohol restrictions and social desirability bias.24,25 The observed associations are biologically plausible. Experimental studies show that intermittent or low-dose alcohol exposure can amplify hepatic oxidative stress, sensitize hepatocytes to lipotoxic injury, and disrupt gut–liver signaling pathways implicated in MASLD progression.26–29 Observational studies in adults similarly demonstrate that moderate drinking accelerates steatohepatitis, fibrosis progression, and incident cirrhosis in individuals with MASLD.30–32 Our findings suggest that early identification and reduction of alcohol use may represent an actionable target for slowing liver injury among youth with MASLD.
Advanced fibrosis at enrollment was associated with lower odds of reporting alcohol use during follow-up, with approximately a 70% reduction compared with participants without fibrosis. This association should be interpreted cautiously because relatively few participants who initiated alcohol use had stage 3–4 fibrosis at enrollment, limiting precision and increasing the possibility that the finding reflects the distribution of disease severity across relatively small alcohol-initiation groups. Potential explanations include more intensive clinical follow-up, greater provider counseling, or heightened personal and family awareness of disease severity, although these mechanisms were not directly measured. Adult studies similarly describe reduced alcohol consumption following the recognition of advanced liver disease, suggesting that perceived vulnerability and provider counseling can shape behavioral choices.6,33 Whether a similar pattern occurs in pediatric MASLD requires further study. In our cohort, higher household income was associated with increased alcohol use, a trend consistent with national data linking socioeconomic advantage to greater adolescent substance exposure. 7 In contrast, alcohol use was not clearly associated with sex, race, or ethnicity, suggesting that clinical and contextual factors may be important determinants of alcohol-associated behavior in youth with MASLD.
This study has several notable strengths. It draws on a large, multicenter, longitudinal cohort of adolescents with biopsy-confirmed MASLD, allowing characterization of baseline disease severity and follow-up across a critical developmental window. The use of standardized NASH CRN protocols ensured uniform assessment of liver chemistries, anthropometrics, and histologic features. Applying group-based trajectory modeling provided new insight into the timing and evolution of alcohol use in youth with chronic liver disease, an area that has received little empirical attention. The study also capitalized on repeated measures of liver injury biomarkers, enabling longitudinal evaluation of liver chemistry trajectories from a common enrollment reference point.
Several limitations should also be acknowledged. Although the overall cohort was large, the trajectory-defined alcohol initiation groups were substantially smaller than the non-use group, limiting statistical power, precision, and the ability to evaluate detailed dose–response relationships or less common higher-intensity drinking patterns. Follow-up duration also differed across alcohol-use groups; although linear mixed models incorporated all available repeated measures and allowed participants with incomplete follow-up to remain in the analysis, unequal follow-up may still have contributed to imprecision or residual bias. Alcohol use was assessed by self-report, which may be subject to recall error and underreporting, particularly given the social sensitivity of drinking behavior in adolescents. Objective biomarkers of alcohol exposure were not available. The cohort did not include systematic repeat imaging or liver biopsy during follow-up, limiting direct assessment of changes in steatosis, inflammation, or fibrosis. Because histology was assessed at enrollment, baseline fibrosis and inflammatory features may not reflect the disease stage later in the follow-up period. Changes in liver chemistries may therefore not fully capture histologic progression. Although models adjusted for key clinical covariates, residual confounding from unmeasured factors such as peer influence, family environment, mental health, diet, and other adolescent risk-taking behaviors, including non-alcohol substance use, cannot be excluded. Site-specific differences in alcohol reporting could not be evaluated reliably because of sparse alcohol-initiation events at some centers. Finally, trajectory groups represent descriptive patterns of behavior rather than causal constructs, and estimates of association should be interpreted in this context.
CONCLUSION
In this multicenter longitudinal cohort of children and adolescents with biopsy-confirmed MASLD, nearly 1 in 5 reported alcohol use during adolescence or early adulthood despite ongoing subspecialty care. Alcohol-use trajectories were associated with greater annual increases in liver chemistries, particularly among participants who reported heavy episodic drinking, supporting the concern that alcohol exposure may contribute to worsening biochemical markers of liver injury during this developmental period. Advanced fibrosis at enrollment was associated with lower odds of alcohol use during follow-up, although the mechanisms underlying this association were not directly measured. These findings emphasize the need for routine alcohol screening and developmentally tailored anticipatory guidance in pediatric MASLD care. Strengthening counseling practices and early prevention strategies may offer a practical opportunity to reduce avoidable liver injury as adolescents with MASLD transition into adulthood.
Supplementary Material
Acknowledgments
FUNDING INFORMATION
The Nonalcoholic Steatohepatitis Clinical Research Network (NASH CRN) is supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (grants U01DK061713, U01DK061718, U01DK061728, U01DK061732, U01DK061734, U01DK061737, U01DK061738, U01DK061730, U24DK061730). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
CONFLICTS OF INTEREST
Jeanne Clark advises Novo Nordisk. Jean Molleston received grants from AbbVie, Albireo/Ipsen, Gilead, and Mirum. The remaining authors have no conflicts to report.
Footnotes
Abbreviations: AASLD, American Association for the Study of Liver Diseases; ALD, alcohol-associated liver disease; ALT, alanine aminotransferase; AST, aspartate aminotransferase; AUDIT-C, Alcohol Use Disorders Identification Test–Consumption; BIC, Bayesian information criterion; BMI, body mass index; BP, blood pressure; CDC, Centers for Disease Control and Prevention; CI, confidence interval; GGT, gamma-glutamyl transferase; GLMM, generalized linear mixed model; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; HED, heavy episodic drinking; HOMA-IR, homeostatic model assessment for insulin resistance; IQR, interquartile range; LDL, low-density lipoprotein; MASLD, metabolic dysfunction–associated steatotic liver disease; NASH CRN, Nonalcoholic Steatohepatitis Clinical Research Network; OR, odds ratio; PNPLA3, patatin-like phospholipase domain-containing protein 3; SD, standard deviation.
Supplemental Digital Content is available for this article. Direct URL citations are provided in the HTML and PDF versions of this article on the journal’s website, www.hepcommjournal.com.
Contributor Information
Nhat Quang N. Thai, Email: nthai@sdsu.edu.
Laura A. Wilson, Email: lwilson9@jhu.edu.
Cynthia A. Behling, Email: cynthiabehlingmd@gmail.com.
Jeanne M. Clark, Email: jeanne.clark@rutgers.edu.
Jean P. Molleston, Email: jpmolles@iu.edu.
Kimberly P. Newton, Email: kpnewton@ucsd.edu.
Rany M. Salem, Email: rsalem@health.ucsd.edu.
Gretchen Bandoli, Email: gbandoli@health.ucsd.edu.
John Alcaraz, Email: jalcaraz@sdsu.edu.
Noe Crespo, Email: ncrespo@sdsu.edu.
Jeffrey B. Schwimmer, Email: jschwimmer@ucsd.edu.
REFERENCES
- 1.Schwimmer JB, Deutsch R, Kahen T, Lavine JE, Stanley C, Behling C. Prevalence of fatty liver in children and adolescents. Pediatrics. 2006;118:1388–1393. [DOI] [PubMed] [Google Scholar]
- 2.Noon SL, Chun LF, Lam TBN, Thai NQN, Dunn W, Schwimmer JB. Prevalence and predictors of suspected metabolic dysfunction-associated steatotic liver disease in adolescents in the United States. Aliment Pharmacol Ther. 2025;61:1479–1488. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Xanthakos SA, Ibrahim SH, Adams K, Kohli R, Sathya P, Sundaram S, et al. AASLD Practice Statement on the evaluation and management of metabolic dysfunction–associated steatotic liver disease in children. Hepatology. 2025;82:1352–1394. [DOI] [PubMed] [Google Scholar]
- 4.Burstein M. Use and abuse of alcohol and illicit drugs in US adolescents: Results of the National Comorbidity Survey–Adolescent Supplement. Arch Gen Psychiatry. 2012;69:390. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Petit G, Kornreich C, Verbanck P, Cimochowska A, Campanella S. Why is adolescence a key period of alcohol initiation and who is prone to develop long-term problem use?: A review of current available data. Socioaffective Neurosci Psychol. 2013;3:21890. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Crabb DW, Im GY, Szabo G, Mellinger JL, Lucey MR. Diagnosis and treatment of alcohol-associated liver diseases: 2019 Practice Guidance from the American Association for the Study of Liver Diseases. Hepatology. 2020;71:306–333. [DOI] [PubMed] [Google Scholar]
- 7.Chapman O, Djerboua M, Rai M, Bechara R, Flemming JA. Alcohol-associated pancreatitis and liver disease among adolescents and young adults. JAMA Netw Open. 2025;8:e2461990. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Liskola J, Haravuori H, Lindberg N, Niemelä S, Karlsson L, Kiviruusu O, et al. AUDIT and AUDIT-C as screening instruments for alcohol problem use in adolescents. Drug Alcohol Depend. 2018;188:266–273. [DOI] [PubMed] [Google Scholar]
- 9.Newton KP, Jayasekera D, Blackford AL, Behling C, Wilson LA, Fishbein MH, et al. Longitudinal response to standard of care in pediatric metabolic dysfunction–associated steatotic liver disease: Rates of improvement and worsening, and factors associated with outcomes. Hepatology. 2025;82:1198–1210. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Newton KP, Lavine JE, Wilson L, Behling C, Vos MB, Molleston JP, et al. Alanine aminotransferase and gamma-glutamyl transpeptidase predict histologic improvement in pediatric nonalcoholic steatohepatitis. Hepatology. 2021;73:937–951. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Li W, Jiang L, Li M, Lin C, Zhu L, Zhao B, et al. Metabolic dysfunction-associated steatotic liver disease is associated with the risk of severe liver fibrosis in pediatric population. Gastroenterol Rep. 2025;13:goaf056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Köylü B, Sökmensüer C, Karçaaltıncaba M, Keskin O. Predicting fibrosis stage in MASH: The role of total metabolic syndrome score and MMP-1. Medicina (Mex). 2025;61:1102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Benyem G, Adu-Gyamfi CO, Afful P, Abotsi GK, Asare KK. Alcohol consumption and its effect on liver function test—A systematic review and meta-analysis. Public Health Environ. 2025;1:17–40. [Google Scholar]
- 14.Wei R, Ogden CL, Parsons VL, Freedman DS, Hales CM. A method for calculating BMI z-scores and percentiles above the 95th percentile of the CDC growth charts. Ann Hum Biol. 2020;47:514–521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.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;41:1313–1321. [DOI] [PubMed] [Google Scholar]
- 16.Juluri R, Vuppalanchi R, Olson J, Ünalp A, Van Natta ML, Cummings OW, et al. Generalizability of the nonalcoholic steatohepatitis Clinical Research Network histologic scoring system for nonalcoholic fatty liver disease. J Clin Gastroenterol. 2011;45:55–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nguena Nguefack HL, Pagé MG, Katz J, Choinière M, Vanasse A, Dorais M, et al. Trajectory modelling techniques useful to epidemiological research: A comparative narrative review of approaches. Clin Epidemiol. 2020;12:1205–1222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jones BL, Nagin DS, Roeder K. A SAS procedure based on mixture models for estimating developmental trajectories. Sociol Methods Res. 2001;29:374–393. [Google Scholar]
- 19.Alexopoulos AS, Crowley MJ, Wang Y, Moylan CA, Guy CD, Henao R, et al. Glycemic control predicts severity of hepatocyte ballooning and hepatic fibrosis in nonalcoholic fatty liver disease. Hepatology. 2021;74:1220–1233. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Volkow ND, Wargo EM. Association of severity of adolescent substance use disorders and long-term outcomes. JAMA Netw Open. 2022;5:e225656. [DOI] [PubMed] [Google Scholar]
- 21.Winters KC, Lee C-YS. Likelihood of developing an alcohol and cannabis use disorder during youth: Association with recent use and age. Drug Alcohol Depend. 2008;92:239–247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.DeWit DJ, Adlaf EM, Offord DR, Ogborne AC. Age at first alcohol use: A risk factor for the development of alcohol disorders. Am J Psychiatry. 2000;157:745–750. [DOI] [PubMed] [Google Scholar]
- 23.US Preventive Services Task Force. Curry SJ, Krist AH, Owens DK, Barry MJ, Caughey AB, et al. Screening and behavioral counseling interventions to reduce unhealthy alcohol use in adolescents and adults: US Preventive Services Task Force Recommendation Statement. JAMA. 2018;320:1899. [DOI] [PubMed] [Google Scholar]
- 24.Koning IM, Harakeh Z, Engels RCME, Vollebergh WAM. A comparison of self-reported alcohol use measures by early adolescents: Questionnaires versus diary. J Subst Use. 2010;15:166–173. [Google Scholar]
- 25.Tourangeau R, Yan T. Sensitive questions in surveys. Psychol Bull. 2007;133:859–883. [DOI] [PubMed] [Google Scholar]
- 26.Buyco DG, Martin J, Jeon S, Hooks R, Lin C, Carr R. Experimental models of metabolic and alcoholic fatty liver disease. World J Gastroenterol. 2021;27:1–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lee J-Y, Jee Y-M, Yang K, Ryu T. Alcohol-induced oxidative stress and gut–liver–brain crosstalk: Expanding the paradigm from ALD to MetALD. Antioxidants. 2025;14:1196. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Dey A, Cederbaum AI. Alcohol and oxidative liver injury. Hepatology. 2006;43:S63–S74. [DOI] [PubMed] [Google Scholar]
- 29.Salete-Granado D, Carbonell C, Puertas-Miranda D, Vega-Rodríguez VJ, García-Macia M, Herrero AB, et al. Autophagy, oxidative stress, and alcoholic liver disease: A systematic review and potential clinical applications. Antioxidants. 2023;12:1425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Jarvis H, O'Keefe H, Craig D, Stow D, Hanratty B, Anstee QM. Does moderate alcohol consumption accelerate the progression of liver disease in NAFLD? A systematic review and narrative synthesis. BMJ Open. 2022;12:e049767. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Parker R, Kim S-J, Gao B. Alcohol, adipose tissue and liver disease: Mechanistic links and clinical considerations. Nat Rev Gastroenterol Hepatol. 2018;15:50–59. [DOI] [PubMed] [Google Scholar]
- 32.Marti-Aguado D, Calleja JL, Vilar-Gomez E, Iruzubieta P, Rodríguez-Duque JC, Del Barrio M, et al. Low-to-moderate alcohol consumption is associated with increased fibrosis in individuals with metabolic dysfunction-associated steatotic liver disease. J Hepatol. 2024;81:930–940. [DOI] [PubMed] [Google Scholar]
- 33.Hemrage S, Brobbin E, Deluca P, Drummond C. Efficacy of psychosocial interventions to reduce alcohol use in comorbid alcohol use disorder and alcohol-related liver disease: A systematic review of randomized controlled trials. Alcohol Alcohol. 2023;58:478–484. [DOI] [PMC free article] [PubMed] [Google Scholar]