A consensus blood transcriptomic framework for sepsis.
Summary
Aggregating MARS and GAinS datasets, the authors define three robust consensus transcriptomic subtypes (CTS1–3) with distinct inflammatory, hemostatic, and interferon/lymphoid signatures. External validation (VANISH RCT cohort and Ugandan cohort) supports generalizability, and post hoc analyses suggest corticosteroids may be harmful in CTS2, highlighting actionable endotyping for trial design.
Key Findings
- Defined three consensus transcriptomic subtypes (CTS1–3) with distinct biological signatures across two major cohorts.
- External validation in VANISH RCT cohort (n=176) and a Ugandan suspected sepsis cohort (n=128) confirmed robustness.
- Post hoc analyses indicated a harmful corticosteroid signal in CTS2-assigned patients.
Clinical Implications
CTS assignment could guide corticosteroid use and other immunomodulators, inform eligibility/enrichment strategies, and enable precision trial designs in sepsis.
Why It Matters
This study provides a standardized, reproducible blood-based endotyping framework, resolving inconsistencies across prior classifications and revealing a potential harmful steroid interaction in a defined subtype.
Limitations
- Treatment interactions were identified post hoc; no prospective stratified randomization to confirm causality.
- Potential batch/platform effects and timing limited to ICU admission samples may affect generalizability.
Future Directions
Prospective, stratified RCTs using CTS assignment to test corticosteroid and other immunomodulatory therapies; development of rapid clinical assays for CTS classification.
Study Information
- Study Type
- Meta-analysis
- Research Domain
- Diagnosis
- Evidence Level
- II - Aggregated analysis of observational cohorts with external validation and RCT data reanalysis; not a randomized intervention.
- Study Design
- OTHER