Metabolome × Microbiome Changes Associated with a Diet-Induced Reduction in Hepatic Fat among Adolescent Boys

**Background:** Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease in children, with no approved pharmacotherapies. Dietary sugar restriction has emerged as a promising therapeutic strategy, but the underlying biological mechanisms remain incompletely understood. Prior cross-sectional studies have identified metabolomic and microbiome disturbances in pediatric NAFLD, but prospective intervention data are lacking. This study aimed to characterize systemic metabolome and microbiome changes associated with a low free-sugar diet (LFSD)-induced reduction in hepatic fat in adolescent boys with NAFLD.

**Methods:** This was an ancillary analysis of an 8-week randomized controlled trial. Forty adolescent boys (mean age 13.0 ± 1.9 years) with biopsy-proven NAFLD were randomized to receive an LFSD (goal: <3% of total energy intake from free sugars; n=20) or consume their usual diet (n=20). Fasting plasma samples were collected at baseline and week 8 for high-resolution metabolomics using dual-platform liquid chromatography-mass spectrometry (HILIC/+ESI and C18/−ESI), yielding 9865 HILIC and 7280 C18 chemical features. After filtering, 6259 HILIC and 4009 C18 features were retained. Differential expression was assessed using baseline-adjusted linear regression models (raw p<0.05), followed by pathway enrichment analysis using Mummichog v2.0.6. In a subsample (n=8 LFSD, n=10 usual diet), 16S rRNA sequencing was performed on stool samples to assess changes in microbial composition and diversity. Metabolite concentrations for selected confirmed (Level 1) metabolites were estimated using reference standardization.

**Key Results:** The LFSD treatment was associated with differential expression of 419 HILIC/+ESI and 205 C18/−ESI metabolite features (raw p<0.05). No features remained significant after FDR correction. Pathway analysis revealed enrichment in amino acid pathways (methionine/cysteine metabolism, serine/glycine/alanine metabolism, tryptophan metabolism, glutamate metabolism) and lipid pathways (omega-3 fatty acid metabolism, linoleate metabolism, bile acid metabolism), as well as vitamin B6 metabolism (all p<0.05). Quantified metabolites that significantly increased in the LFSD vs. usual diet group included serine (p=0.001), glycine (p=0.004), 2-aminobutyric acid (p=0.012), and 3-hydroxybutyric acid (p=0.005). Linolenic acid significantly decreased (p=0.006). Tryptophan-related metabolites showed decreased kynurenine and increased indole-3-acetic acid (p<0.05). For the microbiome, richness increased at the phylum level in the LFSD vs. usual diet group (p=0.02). Two genera increased significantly: an unclassified genus from family Ruminococcaceae (p=0.006) and Phascolarctobacterium from family Veillonellaceae (p=0.035). Ruminococcus bromii also increased (p=0.026). No microbiome findings survived FDR correction. Exploratory integrative analysis revealed correlations between differentially changed metabolites and genera, including a negative correlation between kynurenine and Phascolarctobacterium (r=−0.52, p=0.006, q=0.14). Post-hoc power analysis estimated that 57 participants per group would be needed for metabolomics and 31 per group for microbiome analyses to achieve >80% power after FDR correction.

**Clinical Implications:** This study provides the first prospective evidence linking dietary sugar restriction to specific metabolomic and microbiome changes in pediatric NAFLD. The observed increases in serine, glycine, and 2-aminobutyric acid suggest improved one-carbon metabolism and glutathione homeostasis, potentially reflecting reduced oxidative stress secondary to hepatic fat reduction. The increase in 3-hydroxybutyric acid may indicate improved metabolic flexibility and ketogenic capacity. The lack of change in branched-chain and aromatic amino acids, despite their known disturbance in NAFLD, suggests these may be less responsive to short-term dietary intervention. The microbiome changes, though subtle, point to potential gut-liver axis involvement. These findings identify candidate response biomarkers for future trials and provide effect size estimates for sample size calculations. Larger, adequately powered studies are needed to validate these findings and determine whether the observed changes are driven by sugar restriction per se, hepatic fat reduction, or both.