Identification of a robust metabolic signature associated with hospital-acquired pneumonia and response to interferon-gamma treatment in critically ill patients.

Summary

In a two-center prospective cohort with longitudinal metabolomics, three fatty acid metabolism-driven trajectories stratified HAP risk (24%, 60%, 78%) and ARDS risk (6%, 16%, 43%). These patterns replicated in an independent RCT dataset and aligned with differential effects of interferon-γ1b on ICU discharge alive, suggesting a pathway to predictive enrichment.

Key Findings

• Three longitudinal metabolic response patterns based on fatty acid metabolism stratified HAP risk at 24%, 60%, and 78%.
• The same patterns stratified ARDS risk at 6%, 16%, and 43% in the discovery cohort.
• In an RCT replication dataset, temporal metabolite courses and HAP associations were similar (18%, 28%, 40%).
• Interferon-γ1b decreased the probability of ICU discharge alive in low-risk patterns and increased it in high-risk patterns.

Clinical Implications

Metabolic phenotyping could inform targeted HAP prevention, early ARDS surveillance, and selection of patients more likely to benefit from interferon-γ. Implementation requires prospective validation and operational assays.

Limitations

• Discovery cohort restricted to brain-injured critically ill patients may limit generalizability
• Observational associations; treatment-response findings are exploratory and not randomized by metabolic phenotype

Future Directions

Prospective, phenotype-stratified interventional trials (e.g., interferon-γ) and external validation across diverse ICU populations; development of bedside assays for rapid metabolic classification.