Metabolic septic shock sub-phenotypes, stability over time and association with clinical outcome.
Summary
Across two septic shock RCT cohorts, k-means clustering of 474 serum metabolites revealed three subphenotypes driven by lipid species, notably lysophospholipids. Persistently low lysophospholipids tracked with higher cytokines and increased mortality, while no baseline subphenotype modified trial treatment effects.
Key Findings
- Three metabolic subphenotypes of septic shock were identified, driven mainly by lipid species, especially lysophospholipids.
- Persistent low-lysophospholipid subphenotypes had higher cytokines and higher mortality (e.g., LeoPARDS OR 3.66 and 2.49; VANISH OR 4.13 and 3.22 vs cluster 1).
- No heterogeneity of treatment effect by baseline metabolic subphenotype for trial interventions.
Clinical Implications
Serial metabolite profiling could flag patients with persistently low lysophospholipids at higher risk, informing monitoring intensity and trial enrichment strategies.
Why It Matters
Identifying metabolically defined, prognostically relevant subgroups advances precision sepsis research and suggests testable therapeutic hypotheses (e.g., lysophospholipid augmentation).
Limitations
- Post hoc clustering with potential overfitting and limited generalizability beyond trial populations.
- No demonstrated treatment effect modification to guide current therapy selection.
Future Directions
Prospective validation, development of parsimonious clinical/metabolite panels for bedside use, and interventional trials targeting lysophospholipid pathways.
Study Information
- Study Type
- Cohort
- Research Domain
- Pathophysiology
- Evidence Level
- II - Secondary analysis of RCT cohorts using metabolomics and machine learning to define prognostic subgroups.
- Study Design
- OTHER