Nomogram-based prediction model for extubation failure in preterm infants with invasive mechanical ventilation

Achieving early and successful extubation was a critical aspect in the respiratory management of preterm infants. This study aims to identify variables for assessment and establish a predictive model to estimate preterm infants who may experience extubation failure following invasive mechanical ventilation. A retrospective analysis was conducted on 265 very low birth weight infants (VLBWIs) with neonatal respiratory distress syndrome (NRDS) who received invasive mechanical ventilation and intratracheal surfactant administration in the neonatal intensive care unit (NICU) of a tertiary hospital. Infants were divided into an extubation failure group and an extubation success group based on whether reintubation was required within 72 h after extubation. Independent risk factors for extubation failure were identified, and a nomogram prediction model was constructed. Additionally, 71 VLBWIs from another tertiary hospital, meeting the same inclusion criteria, were used as an external validation dataset to assess the model’s performance. Gestational age, birth weight, patent ductus arteriosus (PDA) diameter, hematocrit (Hct) before extubation, and sepsis were identified as independent risk factors for extubation failure. The nomogram model based on these factors achieved a concordance index of 0.888 for internal validation (training set) and 0.862 for external validation (independent dataset), indicating robust consistency between predicted and observed probabilities. The sensitivity and specificity of the model were 89.8% and 77.8%, respectively, with an area under the receiver operating characteristic curve (AUC) of 0.888 (95% CI: 0.845–0.931). Calibration curves demonstrated a high degree of agreement between predicted and actual probabilities. Gestational age, birth weight, PDA diameter, hematocrit, and sepsis are key factors influencing initial extubation failure in preterm infants. The nomogram model based on these factors demonstrates excellent predictive accuracy and provides valuable guidance for clinical decision-making regarding extubation.