A longitudinal plasma metabolomics dataset of sepsis-associated liver injury across multiple clinical time points

Sepsis-associated liver injury is a severe complication of sepsis characterized by substantial metabolic alterations, yet longitudinal metabolomics resources for this condition remain limited. Here, we present an LC-MS-based longitudinal metabolomics dataset comprising 88 blood samples collected from healthy controls and sepsis patients with or without liver injury across multiple ICU-associated clinical time points. Comprehensive metabolomic profiling generated 18,247 high-quality metabolic features and 2,894 annotated metabolites, providing a resource for investigating temporal metabolite patterns with annotations related to nucleotide, lipid, and bile acid metabolism across clinical groups and sampling time points. Exploratory analytical approaches, including temporal clustering, weighted correlation network analysis, and regularized regression-based feature selection, were applied to facilitate characterization of longitudinal metabolic patterns within the dataset. This dataset provides a reusable resource for future studies on longitudinal metabolic dynamics, biomarker validation, integrative multi-omics analyses, and methodological development in sepsis and critical care research.