RESUMO
OBJECTIVE: This study evaluates the ROX index's accuracy in predicting the success or failure of high-flow nasal cannula (HFNC) therapy in children under 2 years with acute respiratory failure (ARF) from lower respiratory tract infections. METHODS: From January 2018 to 2021 we conducted this multicenter retrospective cohort study, which included patients aged 2-24 months. We aimed to assess HFNC therapy outcomes as either success or failure. The analysis covered patient demographics, diagnoses, vital signs, and ROX index values at intervals from 0 to 48 h after initiating HFNC. We used bivariate analysis, repeated measures ANOVA, multivariate logistic regression, and the area under the receiver operating characteristic (AUC-ROC) curve for statistical analysis. RESULTS: The study involved 529 patients from six centers, with 198 females (37%) and a median age of 9 months (IQR: 3-15 months). HFNC therapy failed in 38% of cases. We observed significant variability in failure rates across different centers and physicians (p < .001). The ROX index was significantly associated with HFNC outcomes at all time points, showing an increasing trend in success cases over time (p < .001), but not in HFNC failure cases. Its predictive ability is limited, with AUC-ROC values ranging from 0.56 at the start to 0.67 at 48 h. CONCLUSION: While the ROX index is associated with HFNC outcomes in children under 2 years, its predictive ability is modest, impacted by significant variability among patients, physicians, and centers. These findings emphasize the need for more reliable predictive tools for HFNC therapy in this patient population.