Accuracy of lung and diaphragm ultrasound in predicting successful extubation in extremely preterm infants: A prospective observational study

Abstract

Objective: Chest ultrasound has emerged as a promising tool in predicting extubation readiness in adults and children, yet its utility in preterm infants is lacking. Our aim was to assess the utility of lung ultrasound severity score (LUSS) and diaphragmatic function in predicting extubation readiness in extremely preterm infants.Study Design: In this prospective cohort study, preterm infants < 28 weeks gestational age (GA) who received invasive mechanical ventilation for ≥12 h were enrolled. Chest ultrasound was performed before extubation. The primary outcome was lung ultrasound accuracy for predicting successful extubation at 3 days.Descriptive statistics and logistic regression were done using SPSS version 22.Results: We enrolled 45 infants, of whom 36 (80%) were successfully extubated.GA and postmenstrual age (PMA) at extubation were significantly higher in the successful group. The LUSS was significantly lower in the successful group compared to failed group (11.9 ± 3.2 vs. 19.1 ± 3.1 p < 0.001). The two groups had no statistically significant difference in diaphragmatic excursion or diaphragmatic thickness fraction. Logistic regression analysis controlling for GA and PMA at extubation showed LUSS was an independent predictor for successful extubation (odd ratio 0.46, 95% confidence interval [0.23-0.9], p = 0.02). The area under the receiver operating characteristic curve was 0.95 (p ˂ 0.001) for LUSS, and a cut-off value of ≥15 had 95% sensitivity and 85% specificity in detecting extubation failure. Conclusion:In extremely preterm infants, lung ultrasound has good accuracy for predicting successful extubation. However, diaphragmatic measurements were not reliable predictors.

Accuracy of lung ultrasound in predicting extubation failure in neonates: A systematic review and meta‐analysis

“…Several meta‐analysis have described the varying diagnostic capabilities of LU in conditions such as transient tachypnea of the newborn, 14–16 respiratory distress syndrome (RDS), 17,18 bronchopulmonary dysplasia, 19 and early surfactant replacement prediction 20,21 . Moreover, recent studies have indicated that LU is a promising tool for prediction of extubation outcomes (success or failure) in neonates 22–29 . These recent findings in neonates are consistent with recent systematic review in adults which concluded that LU can assist in determining the ideal time for extubation 30 .…”

“…While six of the included studies were fully published, 22,23,25,27–29 the remaining two were conference abstracts 24,26 for which the authors provided additional data upon request. Of the eight studies, four included both preterm and term infants, while the remaining four studies focused on preterm infants 24,27–29 . Table 2 provides a summary of the LU methodology of the included studies.…”

“…Quantitative assessment of LU scans, in which a score was given, was conducted in seven studies, whereas one study performed a qualitative assessment 25 . The quantitative scores varied across studies, with a total of 18 points in two studies, 23,24 24 points in one study, 29 36 points in three studies, 22,26,27 and 70 points in one study 28 . Additionally, the cut‐off points for selected sensitivity and specificity were variable.…”

“…Prespecified subgroup analysis was performed for studies using LUS of anterior and lateral six zones (four studies, N = 198), 23–25,29 which showed pooled sensitivity of 0.83 (95% CI: 0.67−0.92), specificity of 0.82 (95% CI: 0.77−0.88), pooled diagnostic odds ratio of 19.32 (95% CI: 7.83−47.68), and AUC 0.91 (95% CI: 0.76−1.0). We observed low statistical heterogeneity (Higgins' I 2 0%, Cochran's Q p value .52).…”

“…Amongst the eight included studies, seven were prospective and one was retrospective in nature. 27 While six of the included studies were fully published, 22,23,25,[27][28][29] the remaining two were conference abstracts 24,26 for which the authors provided additional data upon request. Of the eight studies, four included both preterm and term infants, while the remaining four studies focused on preterm infants.…”