The Diagnostic Utility of Host RNA Biosignatures in Adult Patients With Sepsis: A Systematic Review and Meta-Analysis.

Summary

Across 117 studies (n=17,469; 132 datasets), host RNA biosignatures achieved pooled AUC 0.86 (95% CI, 0.84–0.88) against combined controls. Diagnostic performance remained strong versus healthy controls (AUC 0.87) and against SIRS controls (AUC 0.84). Models with excellent discrimination against SIRS included UrSepsisModel (210 DEGs), microRNA-143, and Septicyte Lab.

Key Findings

• Pooled AUC for sepsis diagnosis using host RNA biosignatures was 0.86 (95% CI, 0.84–0.88) across 117 studies (n=17,469).
• Performance remained robust versus healthy controls (AUC 0.87) and SIRS controls (AUC 0.84).
• UrSepsisModel (210 DEGs), microRNA-143, and Septicyte Lab showed excellent discrimination against SIRS controls.

Clinical Implications

Supports clinical evaluation of RNA-based diagnostics (e.g., Septicyte Lab) to accelerate early sepsis recognition and differentiation from SIRS, while underscoring the need for prospective validation and implementation studies.

Limitations

• Heterogeneity in study designs, platforms, and comparator groups (healthy vs SIRS)
• Limited prospective, real-world clinical validation and potential publication bias

Future Directions

Standardize biosignature panels and workflows; conduct prospective multicenter trials for clinical validation, utility, and cost-effectiveness in routine pathways.