To evaluate whether a model of a daily fitness checklist for spontaneous breathing tests is able to identify predictive variables of extubation failure in pediatric patients admitted to a Brazilian intensive care unit. This was a single-center, cross-sectional study with prospective data collection. The checklist model comprised 20 items and was applied to assess the ability to perform spontaneous breathing tests. The sample consisted of 126 pediatric patients (85 males (67.5%)) on invasive mechanical ventilation, for whom 1,217 daily assessments were applied at the bedside. The weighted total score of the prediction model showed the highest discriminatory power for the spontaneous breathing test, with sensitivity and specificity indices for fitness failure of 89.7% or success of 84.6%. The cutoff point suggested by the checklist was 8, with a probability of extubation failure less than 5%. Failure increased progressively with increasing score, with a maximum probability of predicting extubation failure of 85%. The extubation failure rate with the use of this model was within what is acceptable in the literature. The daily checklist model for the spontaneous breathing test was able to identify predictive variables of failure in the extubation process in pediatric patients.
Objective: This study aimed to identify and evaluate clinical predictors of High Flow Nasal Canulla (HFNC) failure in pediatric patients with acute respiratory failure (ARF). Methods: We conducted a prospective observational cohort study involving 107 children aged 28 days to 14 years admitted to the Pediatric Units of a tertiary hospital. Patients with ARF who received HFNC as respiratory support were included. Clinical and respiratory parameters were monitored at multiple time points. Results: 97 patients were eligible for analysis. The cohort consisted of 76 successes (78.4%) and 21 failures (21.6%). Within the first hour of therapy, a heart rate exceeding 167.5 beats per minute demonstrated particularly high specificity (93.4%) for subsequent failure. By the second hour, two indicators became apparent: a respiratory rate above 47.5 breaths per minute (AUC 0.67, 95% CI 0.58-0.76) and a Respiratory Rate-Oxygenation (ROX) index below 5.8, the latter showing excellent negative predictive value (92%) that could reassure clinicians about continued HFNC suitability. The most powerful predictor emerged at the 6-hour mark, where a ROX index below 4.8 achieved the highest discriminative capacity (AUC 0.73, 95% CI 0.65-0.81) with 91% sensitivity, while the flow-corrected S/F/RRSD ratio below 136 provided complementary specificity (82%). Conclusion: This stud y suggests that early, safe, and effective prediction of HFNC failure in pediatric ARF is achievable, potentially optimizing clinical decision-making. The ROX index (particularly at 6 hours) was the strongest predictor, showing the highest accuracy, sensitivity, and specificity, followed by measurements at 2 and 12 hours.