Pathogen-specific host response in critically ill patients with blood stream infections: a nested case-control study.

Summary

In a rigorously designed, multi-platform transcriptomic study of 341 critically ill patients, the causative pathogen accounted for 41.8% of host blood transcriptomic variance in bloodstream infection. Streptococcal BSI elicited the strongest innate/adaptive signatures and an 8-gene classifier generalized across Streptococcus species and external cohorts, while E. coli and S. aureus BSIs showed distinct cytokine/systemic inflammation and endothelial activation profiles, respectively.

Key Findings

• The causative pathogen explained 41.8% of host blood transcriptomic variance in BSI.
• Streptococcal BSI showed the strongest innate and adaptive immune activation and supported an 8-gene classifier validated across species and external cohorts.
• E. coli BSI aligned with the strongest cytokine/systemic inflammation signals, whereas S. aureus BSI showed the strongest endothelial activation signatures.

Clinical Implications

Transcriptomic classifiers and targeted biomarker panels could enable early pathogen-class inference to tailor antimicrobial choices, timing of source control, and adjunctive therapies (e.g., endothelial stabilizers in S. aureus BSI).

Limitations

• Observational design with sampling within ±1 day of culture may be influenced by early treatments and disease trajectory
• Generalizability to non-ICU or polymicrobial infections requires further study; clinical utility of classifiers needs prospective testing

Future Directions

Prospective, real-time testing of pathogen-classifier-guided management to assess impact on time-to-appropriate therapy, source control, and outcomes; expansion to polymicrobial and fungal BSIs.