Abstract
Objective: To investigate the predictive value of lung ultrasound score (LUS) in the extubation failure from mechanical ventilation (MV) among premature infants with neonatal respiratory distress syndrome (RDS). Methods: The retrospective cohort study was conducted with a total of 314 RDS newborns who received MV support for over 24 h. After extubation from MV, infants were divided into extubation success and extubation failure groups. Extubation failure was defined as re-intubation within 48 h after extubation. Univariate and multivariate logistic regression analyses were used to identify the predictors of the extubation failure. The predictive effectiveness of the combined model and LUS in the extubation failure was assessed by receiver operating characteristic curve, area under curve (AUC), and internal validation. Results: 106 infants failed extubation from MV. The combined model for predicting the extubation failure was performed according to the predictors of gestational age, body length, birth weight, and LUS. The AUC of this combined model was 0.871 (sensitivity: 86.67%, specificity: 74.31%). The AUC of LUS was 0.858 (sensitivity: 84.00%, specificity: 80.69%), and the cutoff value was 18. There was no statistical difference in the predictive power between the combined model and LUS (Z = 0.880, P = 0.379). The internal validation result showed that the AUC of LUS was 0.855. Conclusions: LUS presented a good ability in predicting the extubation failure among RDS newborns after MV.