Introduction: Sepsis, a dysregulated host response to infection leading to organ dysfunction, remains a leading cause of ICU mortality. Early diagnosis and accurate prognosis are crucial for effective management. This study integrates multi-omics data to elucidate the roles of immunity and gut microbiota in sepsis.
Methods: We enrolled 31 septic patients, 25 non-septic ICU patients, and 18 healthy controls. Multiparametric flow cytometry and metagenomic sequencing were performed on blood and fecal samples. DE-SWAN identified critical thresholds for sepsis severity. Prognostic models were developed using random forest algorithms, incorporating immune markers and clinical scores.
Results: For diagnosis, septic patients exhibited alterations in innate immunity, with decreased proportions of NKp46+ NK cells (2.58% vs. 5.35%), effector NK cells (1.32% vs. 2.81%), and early NK cells (0.28% vs. 0.71%), as well as reduced HLA-DR expression on pDCs (MFI: 1,785 vs. 2,854), distinguishing sepsis from non-sepsis (AUC=0.861, 0.849, and 0.849, respectively). Septic patients also had decreased mDCs (0.24% vs. 0.48%), pDCs (0.02% vs. 0.06%), and NK cells (2.58% vs. 5.35%). For prognosis, changes in adaptive immunity were critical, with decreased T cells (43.8% vs. 63.2%), increased CD8+ TN (42.3% vs. 28.7%), decreased CD8+ TE (12.1% vs. 28.6%), and upregulated PD-1 (MFI: 2,476 vs. 1,358) on CD8+ T cells. DE-SWAN identified an APACHE II score of 21 as a threshold for the most significant immune perturbations, with severe sepsis patients exhibiting higher mortality (75% vs. 26%) and septic shock incidence (83% vs. 21%). A prognostic model integrating immune indicators, particularly CD4+ TEM cells, with clinical scores achieved an AUC of 0.94 in predicting outcomes, outperforming models based on clinical scores alone (AUC=0.7). Metagenomic analysis identified B. salyersiae and E. faecium as more prevalent in septic patients, with B. salyersiae associated with alterations in NKp46+ NK cells and HLA-DR on pDCs.
Conclusions: This study identifies potential diagnostic and prognostic markers for sepsis. Dysregulation of the pDC/NK cell axis, T cell remodeling, and distinct metagenomic features contribute to sepsis pathogenesis and severity. Integrating multi-omics data can refine sepsis management approaches.
