Predictors of failure of high flow nasal cannula oxygen therapy in children: a prospective observational study.

Use of heated humidified high flow nasal cannula (HFNC) has become the standard treatment modality for children with acute respiratory distress. This study aimed to identify the factors predicting HFNC failure. Methodology: Children 1 month-12 years of age requiring HFNC support were included in this prospective observational study. Vital signs, e.g. heart rate, respiratory rate (RR), Respiratory Clinical Score (RCS), indices, e.g. SF ratio (SpO2/FiO2) and ROX index (SF/RR) and patient outcomes regarding HFNC success or failure were recorded. Receiver operating characteristics (ROC) analysis with area under the curve (AUC) was done at different time points to find the best cut-off value for SF ratio and ROX index as predictors of HFNC failure. Results: Sixty children requiring HFNC were enrolled. Initial response in the RCS score occurred in 46 patients at 4 hours (interquartile range [IQR]: 2-6) and in RR 6 hours (IQR: 4-12). All patients except one (45/46) who had shown the initial response finally responded to HFNC therapy (75%), and the remaining 15/60 (25%) required escalation to invasive ventilation. Best cut-off value of the SF ratio was ≤192 at 12 hours, with 84.4% sensitivity and 84.6% specificity (AUC, 0.819). Similarly, the best ROX index value to predict HFNC failure was ≤5.27 at 24 hours with 72.1% sensitivity and 88.9% specificity (AUROC 0.864). Conclusion: SF ratio at 12 hours and ROX index at 24 hours are best for predicting HFNC failure in children. Initial response in the RCS score (20% reduction from baseline) by 4 hours is also a good indicator of HFNC success, as none (except one) of these patients failed HFNC therapy.