Immune-Based Biomarkers as Predictors of Mortality in ECMO Therapy for Severe COVID-19 ARDS: Insights from a Retrospective Study.
Rosalia Busà, Giovanna Panarello, Alessia Gallo, Vitale Miceli, Salvatore Castelbuono, Maria Concetta Sorrentino, Giandomenico Amico, Claudia Carcione, Giovanna Russelli, Nicola Cuscino, Monica Miele, Francesca Timoneri, Mariangela Di Bella, Giovanni Zito, Floriana Barbera
Abstract
Open AccessExtracorporeal membrane oxygenation (ECMO) is a vital intervention for patients with severe respiratory failure, particularly in unresponsive acute respiratory distress syndrome (ARDS) cases. However, patient selection for ECMO remains a significant challenge. This study aims to identify novel immune-based biomarkers to improve eligibility assessment and predict outcomes in critically ill COVID-19 patients undergoing ECMO. This monocentric observational retrospective cohort study included 80 patients with severe COVID-19-related pneumonia who required ECMO support due to unresponsive ARDS. The patients were admitted to the intensive care unit (ICU) of IRCCS-ISMETT Hospital between September 2020 and April 2021, before the availability of COVID-19 vaccines. All patients were infected with the original SARS-CoV-2 Wuhan strain. Using machine learning approaches, the study analyzed clinical and laboratory data, cytokine levels, RNA sequencing (RNA-seq), and immune cell profiles collected within two days of hospitalization. The analysis identified a 5.56-fold increased mortality risk in patients presenting with a combination of immune factors: a T cell exhaustion profile, low interferon-alpha (IFNα) levels, and high calprotectin levels. These immune markers were strongly associated with poorer outcomes in patients undergoing ECMO. Our findings highlight the critical role of immune profiling in ECMO patient selection and outcome prediction. Incorporating immune-based biomarkers into clinical assessments may enhance the evaluation of ECMO eligibility and guide treatment decisions, ultimately improving patient outcomes.