Exploratory Analysis of Bedside Variables Associated with Transition Failure in a Selected Cohort of Pediatric Patients Stepped Down from Non-Invasive Ventilation to High-Flow Nasal Cannula After Planned Extubation

Background/Objectives: High-flow nasal cannula (HFNC) is increasingly used after non-invasive ventilation (NIV) for post-extubation respiratory support in children, but evidence to guide the NIV-to-HFNC step-down decision is limited and the decision itself is heterogeneous across centers. In a hypothesis-generating analysis, we aimed to describe transition failure and explore bedside variables associated with it in a physician-selected cohort of pediatric patients stepped down from NIV to HFNC. Methods: This single-center retrospective study included 104 consecutive children (1 month–18 years) extubated, supported with continuous NIV, and stepped down to HFNC. Transition failure was defined as reintubation or re-escalation to NIV within 48 h. Step-down eligibility followed five predefined minimum criteria (Glasgow Coma Scale ≥ 13, FiO2 ≤ 50%, SpO2 ≥ 90%, hemodynamic stability, and absence of acute hypercapnia); re-escalation and reintubation followed clinical criteria routinely applied in our unit, with the precise timing of each decision left to the attending team. A prespecified exploratory multivariable logistic regression model included the ROX index at transition, the duration of NIV, and the PRISM-III score. Internal validity was assessed by bootstrap optimism-correction, five-fold cross-validation, and leave-one-out cross-validation, with a calibration plot. Results: Transition failure occurred in 24 patients (23.1%), with 79.2% within the first 24 h. The ROX index at transition showed the highest univariate discrimination (AUC 0.960, 95% CI 0.916–0.993; cut-off ≤6.0, sensitivity 83.3%, specificity 96.2%). In the exploratory multivariable model, a lower ROX index (adjusted OR 0.10, 95% CI 0.02–0.38; p < 0.001) and a longer NIV duration (adjusted OR 1.12 per hour, 95% CI 1.03–1.23; p = 0.012) were associated with transition failure; NIV duration likely behaves as a marker of unresolved respiratory pathology rather than a causal risk factor. PICU length of stay, pneumonia, and 28-day mortality were higher in the failure group. Conclusions: A transition-moment ROX index ≤ 6.0 and a longer preceding NIV duration were associated with higher risk of step-down failure. These findings are strictly hypothesis-generating, subject to confounding by indication and model optimism, and should not be translated into clinical thresholds before prospective multicenter external validation with pre-specified de-escalation and escalation criteria.