Background
Bronchopulmonary dysplasia (BPD) is a major cause of mortality and morbidity in preterm infants [
1]. Neonatal care has improved dramatically over recent decades, and antenatal corticosteroid treatment, surfactant replacement therapy and more sophisticated assisted ventilation have greatly reduced the severity of respiratory morbidity and mortality. Nonetheless, the incidence of BPD has not decreased [
2‐
5]. Approximately 20 to 40% of very low birth weight (VLBW, birth weight < 1500 g) infants and 12 to 13% of preterm infants born before 32 weeks postmenstrual age (PMA) are affected by BPD [
6,
7].
Identification of perinatal or early postnatal predictors of BPD may enable us to implement potentially preventive strategies before BPD has been established. Numerous studies have attempted to identify predictors for BPD. Scoring systems for predicting BPD that include gestational age, birth weight, gender, patent ductus arteriosus (PDA), sepsis and mechanical ventilation in regression models have been presented [
8]. Laughon et al. [
9] developed a predictive model using data from the NICHD Neonatal Research Network Benchmarking Trial. This model incorporated gestational age, birth weight, race, ethnicity, gender, respiratory support and fractional concentration of inspired oxygen (FiO
2) on postnatal days 7, 14, 21 and 28.
Several studies included chest radiographs as a component of the prediction scoring system for BPD. Chest radiographs are a daily routine examination that is part of intensive care for preterm infants, especially during the early postnatal period. In a study, the day 7 chest radiographs with cystic elements or interstitial changes were associated with oxygen dependency at day 28 [
10]. Greenough et al. [
11] suggested that the day 7 chest radiographs with interstitial shadows or cystic elements predict BPD or death before discharge. Similarly, in our experience, interstitial pneumonia patterns on chest radiographs around the end of the first week of life typically heralded later development of BPD. Furthermore, interstitial pneumonia patterns on the day 7 chest radiographs seem to be associated with histologic chorioamnionitis. The purpose of the present study was to evaluate whether interstitial pneumonia patterns on the day 7 chest radiographs can predict BPD or death and are associated with histologic chorioamnionitis.
Discussion
To date, not a few studies have suggested chest radiograph scoring systems to predict BPD [
10,
11,
14‐
19]. However, the currently described scoring systems are often too complex for clinicians to use at the bedside [
10,
11,
14] or are applied to chest radiographs taken at 3–4 weeks of age when BPD is already being established [
15]. High-resolution chest computed tomography (CT) has been used in recent studies to evaluate the severity of BPD in older-aged infants [
20]. However, it is difficult to apply chest CT in practice to newborn preterm infants for the early prediction of BPD. By contrast, our chest radiograph grading system is very simple in that it only evaluates interstitial changes and consists of only four grades. The ease and clarity of our grading system are supported by the fact that there was a high degree of agreement in the grading results between the two radiologists. Our grading system also uses chest radiographs taken on day 7 of life, which is a time point when whether the preterm infants will progress to BPD or not cannot be judged easily. In this regard, our chest radiograph grading system for the prediction of BPD has advantages over other scoring systems using chest radiographs and is expected to be readily available to clinicians at the bedside.
The present study showed that the interstitial pneumonia pattern on the day 7 chest radiograph was significantly associated with BPD or death before 36 weeks PMA and predicted BPD or death with 98% specificity in VLBW infants who were born before 32 weeks PMA. Accordingly, the interstitial pneumonia pattern on the day 7 chest radiograph can help clinicians identify high-risk group of infants for BPD or death at a relative early postnatal age.
Early identification of high-risk infants for BPD is crucial to implement potentially preventive or early therapeutic interventions for BPD. Aggressive application of non-invasive ventilation, more restrictive use of oxygen, and postnatal corticosteroids or novel therapies that are currently being evaluated, including stem cells, would be more effective when implemented as early as possible before structural damage is solidified [
21‐
26]. The previously suggested prediction models for BPD include birth weight, gestational age, gender, and duration of oxygen therapy and mechanical ventilation as the major risk factors [
8,
9,
27‐
29]. Similarly, birth weight, gestational age, and invasivel ventilation on day 7 of life were significant risk factors for BPD or death in our study, and their effects on BPD or death were adjusted during the multivariate analysis step.
As mentioned above, chest radiographs have also been tested for predictors for BPD [
10,
11,
14‐
19]. Several chest radiographic scoring systems have been suggested as prediction tools for BPD. Yuksel et al. showed that radiographic appearance at 7 days of life reached 71% sensitivity and 88% specificity in predicting BPD [
10]. Chest radiographs were scored according to the volume of the thorax, presence of opacification, haziness, interstitial changes and cystic elements. Greenough et al. [
11] suggested that chest radiographs at 7 days of life can be used to predict BPD or death before discharge. They found that the development of BPD can be predicted by the presence of interstitial shadows or fibrosis, whereas death can be predicted by the presence of cysts. In our study, we focused on the interstitial changes alone in evaluating chest radiographs. The interstitial changes on the day 7 chest radiographs were easily noticeable, and a high degree of agreement between observers was obtained [
16]. Furthermore, interstitial changes on chest radiograph are less influenced by inflation status, which will be variable in the preterm infants. Interstitial changes on chest radiograph indicate lung injury, including airway and pulmonary vascular damages and interstitial edema [
17]. The presence of interstitial changes on chest radiograph is significantly associated with BPD [
10,
11,
18] or adverse respiratory outcomes [
19]. Cystic changes on chest radiograph are also associated with BPD or death [
11]. However, we did not evaluate cystic changes in this study because cystic changes on chest radiographs were more difficult to distinguish from other diseases, such as infection, sequelae of pulmonary hemorrhage, and pulmonary interstitial emphysema [
16]. In addition, cystic changes on chest radiograph represent chronic pathological changes and appear later [
17].
Of note, interstitial changes appear from as early as the first several days of life. By day 7 of life, interstitial changes became obvious in preterm infants with the interstitial pneumonia patterns on their chest radiographs. With regard to predicting BPD using chest radiographic findings, the subjectivity of the reviewers and lack of inter-observer agreement have been reported to be problematic [
30]. In the present study, we used a simple chest radiographic grading system that exclusively assessed interstitial changes, and good inter-observer reproducibility was obtained between two radiologists. Taken together, the interstitial pneumonia pattern on the day 7 chest radiograph may be an early and objective predictor of BPD or death that can be utilized to select preterm infants at high risk of BPD or death as early as possible.
Whereas the interstitial pneumonia pattern on the day 7 chest radiograph was associated with future BPD or death, it was also associated with histologic chorioamnionitis in the past. The association of the interstitial pneumonia pattern on the day 7 chest radiograph with previous histologic chorioamnionitis remained significant after adjustment for birth weight, gestational age, and premature rupture of membrane. Several studies have demonstrated that chorioamnionitis or intrauterine infection can disrupt lung development by inducing fetal pulmonary inflammatory responses [
31‐
35]. A meta-analysis also demonstrated that histologic chorioamnionitis is significantly associated with the increased incidence of BPD [
36]. In this regard, the interstitial pneumonia pattern on the day 7 chest radiographs might correspond to a phase in the pathophysiology of pulmonary inflammation that starts with chorioamnionitis and leads to a later diagnosis of BPD. The most frequently discovered microorganism in the amniotic fluid of pregnant woman with histologic chorioamnionitis is
Ureaplasma urealyticum, which is a proven cause of fetal and neonatal pneumonia [
37]. Thus, the interstitial pneumonia pattern on the day 7 chest radiograph may be a reflection of interstitial pneumonia caused by microorganisms, including
Ureaplasma urealyticum. However, interstitial thickening due to interstitial hypercellularity and focal deposition of elastin and collagen are common pathologic changes revealed in the evolving BPD [
38]. The interstitial pneumonia pattern may also be a reflection of these pathologic changes. At this point, it is not clear whether the interstitial pneumonia pattern is a reflection of interstitial pneumonia caused by microorganisms or a reflection of early pathologic changes observed in evolving BPD.
The main limitation of the present study is the retrospective design from a single center. Further prospective studies are required to validate our results. In addition, further research is needed to clarify the underlying pathology reflected by the interstitial pneumonia pattern on the chest radiograph in preterm infants who later develop BPD.