Background
Lung ultrasound (LUS) has become a widely used bedside examination
technique in neonatal intensive care units (NICUs) because it is
radiationless and can be easily and immediately performed by frontline
neonatologists. A comprehensive and standardized LUS guideline has been
developed1-2 and validated by other
studies3-5. Most neonatal lung diseases can be
diagnosed using LUS, including respiratory distress syndrome
(RDS)6, transient tachypnea of the neonate
(TTN)7, meconium aspiration syndrome
(MAS)8, and pneumothorax (PTX)9.
Some imaging patterns, such as ”compact B-line”, ”white lung” and
”consolidation”, are considered to relate to those diseases.
The most common neonatal respiratory condition in late preterm infants
is TTN, and these infants usually have good outcomes with continuous
positive airway pressure (CPAP) treatment or hood oxygen
support10. However, some degree of surfactant damage,
which can cause secondary RDS11, may occur in severe
or long-lasting TTN. Some late preterm and term infants with RDS also
seem to have a more unfavorable prognosis even if pulmonary
surfactant(PS) is applied12-13. In addition, other
issues (e.g., MAS, PTX, pneumonia) that may lead to severe outcomes are
common in these infants and may only manifest immediately after birth.
Thus, identifying these potential patients is important for
neonatologists. As it is radiationless and convenient, LUS a promising
predictive tool to realize this goal.
Roselyne Brat et al14. described the usefulness of LUS
in predicting PS in preterm infants. They used a relatively precise
scoring system and tested its relation to oxygenation. Others performed
similar research and confirmed Brat’s findings15-16.
However, they did not provide information to predict other respiratory
support needs, and calculating scores according to different images may
be complicated or challenging in some cases. By contrast, Raimondi et
al17-18 used only three straightforward LUS patterns
to predict NICU admission or the need for intubation. Nevertheless, we
believe that using a semiquantitative method would be more precise to
predict respiratory support needs.
Our goal was to test whether a simplified semiquantitative evaluation
method based on high-risk LUS imaging patterns could predict respiratory
support needs. We hypothesized that the number of scanning regions with
”high-risk” patterns has high predictive value for respiratory support
needs in preterm and term infants and is more reliable than assessments
based on respiratory symptoms.