Biological subphenotypes in severe acute hypoxaemic respiratory failure and acute respiratory distress syndrome using rapid prospective classification (SPARC) in the USA: a multicentre, observational, study.

Summary

In a 17-hospital prospective cohort (n=338), rapid plasma-based classification into ARDS/AHRF biological subphenotypes was feasible, with 74% successfully subphenotyped and a median 2.2 h turnaround. The hyperinflammatory ARDS subphenotype (29%) was associated with worse outcomes, supporting readiness for real-time precision trials.

Key Findings

• 338 patients enrolled across 17 hospitals; 74% (250/338) successfully subphenotyped using fresh plasma.
• Median time to subphenotype assignment was 2.2 h overall and 1.9 h among successfully subphenotyped patients.
• Hyperinflammatory subphenotype identified in 29% of ARDS and 23% of severe AHRF, associated with worse clinical outcomes.
• Feasibility improved over time: success rose from 59% in the first 100 to 82% in the last 100 participants.

Clinical Implications

Enables biomarker-based enrichment and risk stratification in ARDS trials, and may inform triage and monitoring once validated broadly.

Limitations

• Observational feasibility study; not designed to test treatment effects.
• Performed within a coordinated network; generalizability to diverse settings requires validation.

Future Directions

Conduct pragmatic precision trials targeting hyperinflammatory ARDS, and externally validate assays and thresholds across broader health systems.