This prospective observational cohort study focuses on developing and validating a predictive model for HFNC failure in patients with SARS-CoV-2-related AHRF, combining clinical parameters and metabolomic profiling. Patients received standard-of-care therapy based on hospital protocols, with clinical data and plasma samples collected at inclusion (T0) for laboratory and metabolomic analyses.
The study consists of two cohorts: a Training Cohort (March 2020 to April 2021) for model development and a Validation Cohort (May 2021 to October 2021) for model validation and external testing. HFNC outcomes were classified as success or failure, with failure defined as requiring orotracheal intubation (OTI). Advanced machine-learning techniques were employed for feature selection and model optimization, targeting precise integration of clinical and metabolomic data.
This model seeks to enhance early identification of HFNC failure risk, enabling timely interventions and personalized management in critically ill AHRF patients.
