Clinical lipidomics in bacterial sepsis: a systematic review of serum and plasma evaluations.

Abstract

Background: Lipid metabolism plays a central role in host-pathogen interactions and immune regulation during bacterial sepsis, with its dysregulation contributing to organ failure and mortality. Blood-based lipidomics has emerged as a promising approach for identifying diagnostic and prognostic biomarkers in sepsis. However, discrepancies in lipid profiles between serum and plasma remain a major challenge for clinical translation. This systematic review synthesizes lipidomic evidence from serum and plasma in bacterial sepsis patients to identify reproducible lipid biomarkers, evaluate methodological heterogeneity and assess associations with clinical outcomes.

Methods: A systematic search of five databases (Scopus, Pubmed, Ovid MEDLINE, Web of Science and Google Scholar) identified nine eligible studies published between 2016 and 2025 that applied untargeted or targeted lipidomics on human serum and plasma samples from adult bacterial sepsis patients (PROSPERO: CRD420251086177).

Results: Across both matrices, reproducible candidate lipid biomarkers included triacylglycerols (TAG), phosphatidylcholines (PC), lysophosphatidylcholines (LPC), sphingomyelins (SM) and ceramides. Synthesized findings revealed characteristically distinct directional trajectories: consistent elevation of TAG, bidirectional shifts of PC, depletion of LPC, bidirectional shifts of SM and elevation of SM. Preliminary trend within the literature indicates that serum-based studies more frequently captured prognostic alterations, whereas plasma-based studies often focused on diagnostic discrimination. Substantial heterogeneity was observed in sample handling, lipid extraction protocols and analytical platforms.

Conclusions: Overall, both matrices offer valuable clinical insights, though these apparent differences in diagnostic or prognostic utility may reflect primary study designs rather than intrinsic matrix superiorities. Methodological heterogeneity limits reproducibility, highlighting the need for standardized workflows to enable translation of sepsis lipid biomarkers.