Discussion
This study demonstrated the ability of point of care LUS to predict success of extubation in mechanically ventilated premature infants. Infants with higher LUS were supported with higher venitlatory pressure. LVEI did not differ among infants who succeeded or failed extubation. However, it correlated with PAP.
Infants with higher LUS were supported with significantly higher pressure of conventional mechanical ventilation. The use of higher PIP is indicative of decreased compliance in infants with severe lung disease. A previous study of CPAP-supported infants demonstrated the need for higher pressure in infants with higher LUS12. Oxygen saturation index (OSI) was also increased in infants with severe lung disease who had higher LUS. This finding agreed with previous studies that showed correlation of LUS with OSI and with the ratio of arterial oxygen saturation to FiO212,15.
This study is the first to use LUS to predict success of extubation from mechanical ventilation in premature infants. Many lung ultrasound scores have been developed to assess lung aeration and guide respiratory care in adults especially those with restrictive lung disorders; LUS in this population is strongly recommended (level of evidence A)18. Compared to adult literature, lung ultrasound has not been adequately addressed in neonates although it is easier owing to the small patients’ size and the absence of obesity or heavy musculature19. A few studies described the ultrasound patterns in common respiratory conditions in neonates, such as meconium aspiration syndrome, hyaline membrane disease, transient tachypnea of the neonate, and pneumothorax 21-23. Other studies described the usefulness of LUS in predicting the need for invasive mechanical ventilation in infants supported with noninvasive ventilation24,25 and the need for surfactant replacement in CPAP-supported extremely preterm neonates 26. However, there is no study to predict success of extubation from mechanical ventilation. Giving the ease of its use at the bedside and the non-invasive nature of LUS, it is helpful to obtain LUS before extubation to avoid the exposure of infants to re-intubation should the extubation attempt fail. One of the main barriers to the more extensive use of the ultrasound technology in premature infants is the lack of efficient training solutions and the need to have structured quality-check assurance 27. Once training is established the use of LUS can be the first-line imaging technique in preterm infants 28.
Evaluating an infant for extubation readiness is clinically a challenge. There are significant variations in the decision making process of extubation; that lacks objective evidence-based criteria and is often contingent on caregiver experience. A recent international survey showed extubation readiness subject to caregiver personal interpretations of blood gas parameters and overall feeling of clinical stability of infants 29. Some physicians advocate conducting apnea test before extubation, although this test is often conducted variably30. Therefore, it is important to have an objective method and/or a scoring system with a calculated prediction of success rate accordingly. LUS fulfils this unmet gap in neonatal practice. In this study, when LUS was ≤11extubation was successful in >90% of infants.
Pulmonary hypertension in newborns (PPHN) is triggered by multiple etiologies including hypoxemia and underlying parenchymal lung diseases and can potentially hinder extubation 28. Performing a full echocardiographic studies on all mechanically ventilated preterm infants would be exhaustive and may not consistently correlate with PPHN14. LVEI is a quantifiable measure of the amount of distortion of ventricular septal geometry due to elevated right ventricular systolic or diastolic pressures and/or volumes. Greater degrees of LVEI have been associated with PPHN in children and adults, but have not been studied thoroughly in premature infants. LVEI has the added benefit of being easily measured from any short axis view of the mid-left ventricle 13. Increased end-diastolic LVEI would indicate volume overload as in hemodynamically significant PDA and increased end-systolic LVEI may indicate right ventricular volume overload as in PPHN 31. In the current study, systolic LVEI correlated with PAP that was measured concomitantly by echocardiography. It correlated with the presence of PDA as well. However, end-systolic LVEI did not correlate with OSI and did not prove to be valuable in predicting readiness for extubation. Of note, in this study we used LVEI only as a surrogate for pulmonary hypertension although a full assessment of PAP may have shown benefit in predicting extubation failure.
This study has the strength of addressing the success of extubation trials that is a real challenge in premature infants. The study has some limitations including the lack of comparison of LUS with findings of MRI or CT scan of the chest. To ensure consistency and reliability of the scores, all LUS studies were performed by the same investigator who was an experienced pediatric radiologist. However, these studies were well within the scope of POCUS and as suggested by international guidelines would be practically performed by neonatal caregivers with minimal experience with US 7,32. Similarly, we have identified the US machine and proble models that were used for this research, although other devices with reasonable resolution will suffice32. LUS addresses failure of extubation related to lung parenchymal disease. Other factors involved in extubation failure such as airway edema and apnea of prematurity cannot be predicted with LUS. Respiratory management in this center allows mechanical ventilation for several days, therefore the investigators advise using caution when extrapolating the current findings in settings that have different respiratory practices. In fact, being ventilated for several days, the respiratory failure for these enrolled infants would be better defined as an early stage of chronic pulmonary insufficiency rather than the classical respiratory distress syndrome 33.