Quantitative lung ultrasound to guide surfactant retreatment in preterm neonates born at ≤30 weeks' gestation: a multicentre retrospective non-inferiority diagnostic accuracy study.

Abstract

The most critically ill neonates may require repeated surfactant doses, but there are no consensus criteria for retreatment. We aim to verify whether lung ultrasound (LUS) aeration score predicts surfactant retreatment with accuracy equal to that known for the first administration. This was a multicentre, retrospective, non-inferiority, diagnostic accuracy study. Preterm (≤30 weeks' gestation) neonates were enrolled and LUS aeration score was calculated at T1 (i.e., ≈1 h of life) and before surfactant administration, if any; and T2: ≈12 h of life and at least 10 h after the first surfactant administration. The area under the curve (AUC) was chosen considering the score as triage (i.e., highest sensitivity) and replacement test (i.e., highest global accuracy). AUC was higher for T2 (AUC = 0.854 (95% CI: 0.84; 0.87), p < 0.0001) than for T1 score (AUC = 0.69 (95% CI: 0.67; 0.71), p < 0.0001; difference = 0.165 (95% CI: 0.116; 0.214), p < 0.001). AUC of T2 score was similar to that previously reported at T1 for prediction of the first surfactant treatment in preterm (p = 0.225) or in late preterm/term patients (p = 0.579). Cut-offs to use the T2 score as triage (score = 4) or replacement (score = 8) test had a sensitivity of 98% and a global accuracy of 78%, respectively. Accuracy is independent of gestational age, and the T2 score is associated with the surfactant retreatment (aOR = 1.57 (95% CI: 1.38; 1.79), p < 0.001). In preterm neonates, LUS aeration score calculated at ≈12 h of life and at least 10 h after the first surfactant administration predicts retreatment with accuracy equal to that of the score calculated at ≈1 h of life to predict the first administration. The accuracy is independent of gestational age. None.