Analyzing big data: could the ROX index predict risk for intubation in a surgical patient receiving NIV/CPAP?

Background and goal of study: The ROX index is defined as the ratio of SpO2 measured by pulse oximetry/FiO2 to respiratory rate. It is used as predictor of the need for intubation in patients receiving HFNC oxygen therapy, but not yet validated for NIV/CPAP ventilation. The goal of this study is to determine if the ROX index could predict risk for intubation in surgical patients receiving NIV/CPAP mask ventilation. Materials and Methods: Patient data was extracted from freely available AmsterdamUMC ICU database containing data from a 32- bed mixed surgical- medical academic ICU with 1 billion clinical data points related to 23106 admissions between 2003 and 2016. We selected patients that were non- invasively mechanically ventilated and extracted their demographic data (age, sex, BMI) and SOFA score at admission, as well as ROX index from 2 hour window mean values of SpO2, RR and FiO2 for the first 12 hours of ventilation. Results and Discussion: From 20106 patients, a cohort of 1075 NIV patients using CPAP/PS modes were identified. We excluded patients that received NIV ventilation after an episode of intubation (n=47) and patients that were intubated more than 24 hours after NIV episode (n=53). Final dataset consisted of 975 patients, 190 were intubated within 24 hours NIV ventilation start. Patient demographic data was similar between intubated and non- intubated group (Table 1). Conclusion(s): Analysing a large database of surgical patients, we found that ROX index significantly correlates with the need for intubation after 4 hours of NIV/CPAP ventilation. ROX index could be a good predictor for intubation in surgical patients receiving NIV CPAP/PS ventilation.