ABSTRACT Objective This study aimed to identify early clinical predictors of non‐invasive ventilation (NIV) failure in children with Pediatric Acute Respiratory Distress Syndrome (PARDS) and to develop a model to predict NIV outcomes. Methods A prospective observational study including children (0−15 years) who were supported by NIV and met PARDS criteria was performed in two public PICUs in Chile. Demographic, physiological variables, and clinical outcomes were recorded. Logistic regression was used to identify risk factors for NIV failure and develop a predictive model. Results Seventy‐six patients were enrolled (median age 22 months IQR 11–48; weight 12.1 kg 9.6–18.3). Eighteen patients (24%) failed NIV (Failure group), with a median NIV duration of 12.5 h (IQR 8–23.8). The Failure group showed a significant reduction in the SpO 2 /FiO 2 ratio after 2 h of NIV (−59 −124 to 5 vs −2 −49 to 30, p = 0.03). In multivariate analysis, an increase in ΔSpO 2 /FiO 2 (improvement in oxygenation) was associated with lower odds of NIV failure (OR 0.98; 95% CI 0.97–0.99; p = 0.006), indicating that worsening oxygenation was associated with a higher risk. The final prediction model included age and early changes in SpO 2 /FiO 2 , respiratory rate, and heart rate, demonstrating good discrimination (AUROC 0.764; 95% CI 0.641–0.888). Conclusion In children with PARDS treated with NIV, a model including age and early changes in oxygenation, heart rate, and respiratory rate within the first 2 h predicts the need for intubation with good discrimination.
Bustos-Gajardo et al. (Wed,) studied this question.