Dynamic oxygenation subgroup bringing new insights in ARDS: more predictive of outcomes and response to PEEP than static PaO

Summary

Across five ARDS datasets (n=814 training; n=2505 validation), three longitudinal PaO2/FiO2 trajectory-based subgroups were derived over the first 3 days after ARDS diagnosis. These dynamic subgroups better predicted prognosis and response to PEEP than static Berlin severity categories.

Key Findings

• Developed three longitudinal PaO2/FiO2 trajectory subgroups over the first 3 days after ARDS diagnosis.
• Validated the subgroups across four external cohorts (FACTT, SAILS, ALVEOLI, MIMIC-IV; total validation n=2505).
• Dynamic subgroups were more predictive of prognosis and PEEP response than static Berlin categories.

Clinical Implications

Clinicians could incorporate the first 72-hour oxygenation trajectory to stratify risk and tailor PEEP strategies, moving beyond static PaO2/FiO2 thresholds. It also provides a framework for trial enrichment and adaptive ventilation strategies.

Limitations

• Details on absolute effect sizes and clinical decision thresholds are not fully reported in the abstract.
• As a secondary analysis, potential confounding and protocol heterogeneity across datasets remain.

Future Directions

Prospective interventional trials stratifying patients by early oxygenation trajectories to test personalized PEEP/ventilation strategies; integration with EIT-derived V/Q metrics.