Predictive Value of ROX and ROX-HR for Care Escalation in Infants With Bronchiolitis Initiated on High-Flow Nasal Cannula

OBJECTIVE: Current scoring tools are suboptimal to predict care escalation in infants with bronchiolitis. Single-center studies suggest that respiratory oxygenation index (ROX) and ROX-heart rate (ROX-HR) may have predictive value. We evaluated the ability of these measures to predict care escalation in a multicenter cohort of infants with bronchiolitis initiated on high-flow nasal cannula (HFNC). METHODS: We performed a secondary analysis of a retrospective multicenter cohort study of infants with bronchiolitis initiated on HFNC at 3 pediatric emergency departments (EDs) between February 1, 2018 and March 1, 2020. ROX [(SpO2/FiO2)/RR] and ROX-HR (ROX/HR × 100) were calculated at triage and pre-HFNC initiation. We defined care escalation as the need for positive pressure ventilation or ICU care. RESULTS: Of 738 infants, 73 (9.9%) required care escalation. These infants had higher maximum heart and respiratory rate, greater proportion were hypoxemic and required higher HFNC support. While there were no significant differences in mean initial ROX [9.4 (3.1) vs 9.4 (3.0); P = 0.81], pre-HFNC ROX [9.4 (3.0) vs 9.0 (2.9); P = 0.24] and initial ROX-HR [6.4 (2.7) vs 6.4 (2.5); P = 0.94], there was a small difference in pre-HFNC ROX-HR (5.7±2.0 vs 5.4±2.0; P = 0.048) between those who did and did not require care escalation. AUROCs for initial and pre-HFNC ROX were 0.51 (95% CI, 0.43-0.58) and 0.54 (95% CI, 0.47-0.61), and initial and pre-HFNC ROX-HR were 0.50 (95% CI, 0.43-0.58) and 0.57 (95% CI, 0.49-0.64), respectively. CONCLUSION: ROX and ROX-HR showed poor ability to predict care escalation in infants with bronchiolitis initiated on HFNC.