Abstract
PURPOSE: Alcohol use disorder (AUD) is a chronic relapsing condition frequently complicated by cognitive impairment (CI), yet its underlying metabolic mechanisms remain poorly understood. This study aimed to identify plasma metabolic signatures and dysregulated pathways associated with AUD-CI using an integrated multi-omics approach. METHODS: A prospective cohort study of 210 male participants (70 AUD-CI, 70 AUD without CI [AUD-NonCI], and 70 healthy controls [HCs]) was conducted. Plasma samples underwent LC-MS/MS-based metabolomic and lipidomic profiling. Cognitive function was assessed using the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). Machine learning algorithms (Random Forest and LASSO regression) were employed for biomarker selection, and pathway analysis was performed using MetaboAnalyst 5.0. RESULTS: The multi-omics platform detected 117 differentially expressed molecules (11 metabolites and 106 lipids) with high diagnostic accuracy (mean AUC=0.92 ± 0.03). Key findings included depletion of S-adenosylmethionine (SAM; 1.8-fold decrease, p=3.4×10(-4)) and accumulation of ceramide species Cer (d18:1/26:2) (2.1-fold increase, p=7.8×10(-4)). Pathway analysis revealed mTORC1 signaling (p=1.4×10(-4)) and sphingolipid metabolism (p=2.1×10(-5)) as central dysregulated pathways. AUD-CI patients exhibited 49 unique lipid alterations, notably 70% reduction of phosphatidylcholine PC (42:4) versus HCs (p=0.002), strongly correlated with synaptic protein markers (r=0.82, p<0.001). CONCLUSION: Our findings characterize a dysregulated liver-gut-brain metabolic axis in AUD-CI pathogenesis, highlighting the mTORC1-sphingolipid pathway as a promising therapeutic target. The identified biomarkers provide mechanistic insights into alcohol-induced neurotoxicity, offering potential avenues for precision interventions in AUD-related cognitive decline.