SpO₂/FiO₂ as a predictor of non-invasive ventilation failure in children with hypoxemic respiratory insufficiency

Generally, monitoring of hypoxemia has been always based on partial pressure of arterial oxygen to fraction of inspired oxygen ratio; recently it was confirmed a good correlation between this ratio and a non-invasive alternative ratio of oxygen saturation to fraction of inspired oxygen (SF ratio = SpO₂/FiO₂). The SF ratio has been studied and validated, especially in the intubated patients, however studies monitoring patients treated with non-invasive ventilation (NIV) are lacking. The aim of the study was to determine if the SF ratio is a reliable predictive factor for failure of NIV in pediatric patients diagnosed with hypoxemic acute respiratory failure due to pneumonia. A five year retrospective study with consecutive sampling of patients diagnosed with hypoxemic acute respiratory failure due to pneumonia, admitted to the pediatric intensive care unit in a university hospital and treated with NIV was conducted. Physiological and ventilator variables were collected before starting NIV and at 2, 8 and 12 hours of treatment, and the corresponding SF ratios calculated. Thirty-four patients were studied. NIV failed in seven patients. Significant differences in SF ratio were observed between the successful and unsuccessful patients at the onset, 2 and 12 hours of treatment. Multivariable analysis revealed the SF ratio at two hours to be an independent predictor of NIV failure (odds ratio = 0.96, 95% confidence intervals 0.93-0.99; P = 0.015). Area Under Receiver Operating Characteristic curve = 0.90 (95% confidence interval = 0.79 to 1.000), optimal cut off value for SF at 2 hours ≤ 189; sensitivity = 86%, specificity = 74%. In conclusion, the SF ratio of 189 or less at 2 hours of NIV seems to be useful for predicting NIV failure in this cohort.