Predictive Model for High-Flow Nasal Cannula Failure Based on Metabolomics and Clinical Data: COVID-19 As a Use Case
Predictive Model for High-Flow Nasal Cannula Failure Based on Metabolomics and Clinical Data: COVID-19 As a Use Case
The aim of this prospective observational cohort study is to develop a predictive model for high-flow nasal cannula (HFNC) failure in patients with acute hypoxemic respiratory failure (AHRF) caused by SARS-CoV-2. By combining clinical data and metabolomic profiles, the study seeks to answer:
- Can integrating clinical parameters with metabolomic markers enhance the prediction of HFNC failure in AHRF patients?
Participants with AHRF due to SARS-CoV-2 receiving HFNC under standard-of-care protocols will be monitored until hospital discharge or death. Blood samples for metabolomic analysis will be collected at inclusion.
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.
Inclusion Criteria:
Exclusion Criteria: