What this Trial Is About

Acute respiratory distress syndrome (ARDS) is a common and life-threatening condition in intensive care units, characterized by substantial biological and clinical heterogeneity. Differences in patients' inflammatory responses, baseline immune function, and organ failure patterns contribute to variability in ARDS severity, treatment response, and clinical outcomes. Precision classification of ARDS based on biological and inflammatory characteristics may therefore be essential for improving patient outcomes. Previous analyses of randomized clinical trials have identified two reproducible inflammatory subphenotypes-"hyperinflammatory" and "hypoinflammatory"-which differ in organ dysfunction profiles, clinical trajectories, and responses to treatments such as fluid management strategies, corticosteroids, and ventilatory interventions. However, key uncertainties remain, including whether these inflammatory subphenotypes can be validated in Chinese ARDS populations, how various bedside prediction models perform in identifying these subphenotypes, and whether model-based subphenotype identification can guide individualized treatment decisions. This multicenter cohort study aims to: (1) validate inflammatory subphenotypes of ARDS using latent class analysis; (2) compare the predictive performance of existing bedside models for subphenotype identification; and (3) assess whether subphenotype assignment based on prediction models can guide individualized treatment strategies, including fluid management, PEEP titration, and corticosteroid use. In addition to these primary aims, the study may include other exploratory objectives, such as evaluating subphenotype stability over time, characterizing biological pathways associated with subphenotypes, and assessing additional treatment-response patterns to support future precision ARDS management strategies.

Validation and Precision Treatment of Inflammatory Subphenotypes in Acute Respiratory Distress Syndrome: A Multicenter Cohort Study