original_data

Background: Blood gas parameters and immunity are associated with the prognosis of septic patients. The aim of this work is to develop a blood gas parameters-based signature (BGBS) that could be used to predict patient prognosis and immune responses.Methods: Medical records retrospectively extracted from the Medical Information Mart for Intensive Care IV were used for constructing BGBS in septic patients and for internal validation in septic shock patients. Bioinformatics analysis and machine learning techniques were adopted to reveal the relationship between BGBS and the primary outcome, 28-day mortality, in sepsis and septic shock.Results: A BGBS containing PO2, BE, and lactate was developed and verified as an independent and reliable risk model that correlated with the 28-day mortality of septic patients (odds ratio: 1.559; 95% confidence interval: 1.464-1.659; P