Prone position ventilation-induced oxygenation improvement as a valuable predictor of survival in patients with acute respiratory distress syndrome: a retrospective observational study

Background In patients with severe acute respiratory distress syndrome (ARDS), prolonged and inappropriate use of prone position ventilation (PPV) is a known risk factor for mortality. Hence, it is critical to monitor patients' response to PPV and accurately differentiate responders from non-responders at an early stage. The study aimed to investigate the relationship between oxygenation improvement after three rounds of PPV and survival rate in patients with pulmonary ARDS. Additionally, we sought to identify the earliest turning point for escalation from PPV to extracorporeal membrane oxygenation. Methods We performed a retrospective observational study from 2015 to 2023. We included adult patients who received invasive mechanical ventilation, underwent at least three periods of at least 6 h of PPV after admission to the Intensive Care Unit, and meet the ARDS criteria. The study collected data on each PPV session, including changes in PaCO2, PaO2, pH, FiO(2), PaO2:FiO(2) ratio, and clinical outcomes. Results A total of 104 patients were enrolled in the study. The change in PaCO2 from baseline to the third PPV session (P3) had the highest area under the receiver operating characteristic curve (AUC) of 0.70 (95% CI 0.60-0.80; p < 0.001) for predicting hospital mortality, with an optimal cut-off point of 3.15 (sensitivity 75.9%, specificity 56.0%). The percentage change in PaO2:FiO(2) ratio from baseline to P3 also had significant AUC of 0.71 (95% CI 0.61-0.81; p < 0.001) for predicting hospital mortality, with an optimal cut-off value of 99.465 (sensitivity 79.6%, specificity 62.0%).PaCO2 responders were defined as those with an increase in PaCO2 of <= 3.15% from baseline to P3, while PaO2:FiO(2) responders were defined as those with an increase in PaO2:FiO2 ratio of >= 99.465% from baseline to P3. In the multivariable Cox analysis, PaO2:FiO(2) responders had a significantly lower 60-day mortality risk (hazard ratio 0.369; 95% CI 0.171-0.798; p = 0.011). Conclusions The percentage change in PaO2:FiO2 ratio from baseline to P3 was a significant predictor of outcomes. The model fit and prediction accuracy were improved by including the variable of PaCO2 responders.