Predicting High-Flow Nasal Cannula Therapy Outcomes Using the ROX-HR Index in the Pediatric ICU

BACKGROUND: High-flow nasal cannula (HFNC) use is increasing in pediatric patients. Objective measures that predict HFNC outcomes are lacking. The respiratory rate-oxygenation (ROX) and ROX heart rate (ROX-HR) indices are validated to predict HFNC therapy failure in adults. This study examined the performance of both indices in predicting HFNC therapy failure in children admitted to the pediatric ICU (PICU). METHODS: This retrospective, longitudinal, observational cohort study was completed in a 24-bed PICU in a quaternary care children's hos-pital. All subjects <= 24 months of age initiated on HFNC in the PICU from January 1, 2018- August 31, 2020, were included. The ROX and ROX-HR indices were collected at standardized time points during HFNC therapy. Performance in predicting HFNC failure was evaluated using area under the receiver operating characteristic curve (AUROC) and Kaplan-Meier survival analysis. Failure was defined as escalation of respiratory support to either noninvasive ventila-tion or endotracheal intubation. RESULTS: Among 446 subject encounters, 111 (24.9%) failed HFNC therapy. HFNC failure was associated with lower ROX and ROX-HR indices at termina-tion compared to HFNC liberation (P < .001). A ROX-HR index < 3 was significantly associated with a higher risk of HFNC failure at 1 (AUROC 0.76, P = .01) and 6 (AUROC 0.81, P = .02) h. CONCLUSIONS: ROX-HR may be a useful tool for early identification of patients <= 24 months at risk for HFNC failure and allow for earlier intervention. Larger prospective studies are necessary to validate the utility of the ROX-HR index in pediatric patients.