Multi-omics signatures of 8-year trajectories and heterogeneity of metabolic dysfunction-associated steatotic liver disease reveal serum citramalic acid as a key contributor

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a heterogeneous condition, yet early predictors for risk stratification remained limited. We aimed to identify multi-omics features predicting MASLD progression and heterogeneity. This population-based cohort study included 1694 participants from the Guangzhou Nutrition and Health Study. Gut metagenomics and targeted fecal and serum metabolomics were profiled. MASLD severity was defined using MRI-derived hepatic steatosis, and 8-year MASLD trajectories were assessed using repeated ultrasound examinations. We identified 9 microbial species, 13 fecal metabolites, and 13 serum metabolites associated with MASLD severity, with serum citramalic acid emerged as the top-ranked feature. For 8-year MASLD trajectories, microbial-, fecal-, and serum metabolite-based risk scores predicted MASLD incidence (per-SD HRs: 1.04, 1.18, and 1.21) and showed stronger associations with persistent MASLD (per-SD ORs: 1.27, 1.62, and 2.23). Multi-omics clustering further revealed three MASLD subtypes with distinct metabolic profiles. Metabolic dysfunction progressively worsened from Subtype A to Subtype C ( P trend < 0.05), with Subtype C exhibiting increased T2D risk (OR: 2.70; 95% CI: 1.61–4.55). Key findings were replicated in an external validation cohort. Furthermore, citramalic acid administration in HFD-fed mice exacerbated metabolic dysfunction and hepatic steatosis, and disrupted hepatic pathways involved in lipid metabolism and cellular stress responses. Our study identified multi-omics features underlying MASLD progression and heterogeneity, providing early biomarkers for risk stratification and potential precision prevention of MASLD.