Bronchopulmonary dysplasia (BPD) is a common chronic respiratory disease in preterm infants (1). The increase in the survival rate of premature babies following the improvement of perinatal treatment and care has caused an increase in the incidence of BPD in recent years (2), which has seriously affected the quality of life of preterm infants. According to the consensus reached at the workshop sponsored by the National Institute of Child Health and Human Development (NICHD) in 2001 (3), BPD was clinically defined based on oxygen dependency in preterm infants. However, the refined NICHD definition of BPD in 2018 (4) emphasized imaging findings to support a diagnosis of lung parenchyma disease.

Fibrotic opacities and cystic changes on chest imaging [chest X-ray (CXR) or computed tomography (CT) scan] were considered typical findings in BPD patients (5–7). In patients with severe BPD, the presence of bubbles/cystic appearance on CXR after 28 days of life was reported to be an important factor, and typical imaging findings can predict a poor pulmonary outcome in BPD patients (8). BPD is associated with poor outcomes (9). Although many studies have been conducted on BPD, there are limited reports specifically evaluating the association of typical imaging findings with clinical characteristics and later outcomes in patients with BPD.

We hypothesized that BPD with typical imaging findings was likely to be a particular subgroup of this entity, with a unique etiology, clinical characteristics and prognosis. Therefore, this retrospective study aimed to compare clinical characteristics, short-term outcomes and follow-up data until 2 years of age in preterm infants with or without typical imaging findings of BPD on CXR or CT scan during the entire hospital stay. A propensity score analysis was used to reduce bias between the two groups, and multivariate logistic regression analysis was performed to identify factors related to mortality in preterm infants with BPD.

BPD is a prevalent respiratory complication of preterm birth. Radiography plays a critical role in the clinical evaluation of BPD (13, 28, 29). Our data demonstrated that typical imaging findings were associated with the severity of BPD, a higher rate of mortality and HOT at discharge. In particular, the typical imaging findings in the early stage (within 7 days) were closely related to mortality.

In the original definition of BPD proposed by pediatric radiologist Northway in 1967 (30), abnormal chest radiographic findings were classified into four stages according to lung pathology. Serial chest radiographs indicated the evolution of BPD. As the survival rate of early preterm infants increased, the radiologic changes were not as classic as Northway proposed because of changed pathophysiology (31). However, neonatologists still attached great importance to imaging confirmation in BPD (32). Fibrotic opacities/cystic appearances in chest radiographs are recognized as the typical radiographic pattern of BPD (33); these appearances correlate with disease severity and are relatively easy to recognize. The typical imaging findings after 36 weeks may not have an impact on the diagnosis and severity of BPD, but they are closely related to mortality and have certain indications for clinical management and treatment.

The clinical ability to reliably predict the duration of respiratory support, let alone the outcomes of infants with BPD, is not yet sufficient. Exploring additional factors related to BPD is the key to improving its treatment and prognosis. Previous studies have reported that the findings of chest radiographs, such as interstitial shadows, fibrosis or bubbly/cystic appearances, could predict the incidence and severity of BPD (34) and have an association with respiratory outcomes. Some infants with moderate or severe BPD may have typical radiographic changes within the first 7 days of life (35). Ren et al. (36) revealed that infants with BPD-related imaging changes required a longer duration of IMV and parenteral nutrition and were more likely to require HOT. However, these studies contained insufficient clinical data and did not adjust for confounding factors. Most of them only paid attention to the chest X-ray appearance within 7 days or after 28 days. In addition, these studies did not analyze long-term prognosis, such as mortality. Therefore, further studies are needed to identify the association between typical imaging findings and outcomes.

The mortality rate of infants with BPD varies among different NICUs. Our study demonstrated that BPD infants with typical imaging findings had a higher probability of developing moderate to severe BPD, as well as a higher mortality rate, than those without typical imaging findings. The overall mortality rate in infants with BPD was 15.2% (39/256). The relatively high mortality rate of BPD is partially due to deficiencies in my country's medical conditions, which leads to delays in individualized and effective treatment.

The multivariate logistic regression analysis showed that typical imaging findings ≤ 7 days and typical imaging findings >7 days were risk factors for mortality in preterm infants with BPD (OR 7.794, p = 0.004; OR 4.533, p = 0.001, respectively). The significant associations we demonstrated between typical imaging findings and the severity of BPD and mortality are in accordance with previous findings. Infants who had undergone IMV ≥ 7 days had a 2.399 (95% CI 1.049–5.484) times higher risk of mortality than those who had undergone IMV <7 days (37), while postnatal glucocorticoid administration and a maternal PIH history were protective factors against mortality (OR 0.201, p < 0.001; OR 0.200, p = 0.041; respectively). This was consistent with other studies (38, 39). Interestingly, the risk factors for mortality in infants with BPD did not include GA or BW. Further work is needed to determine the clinical reason for this result.

Our study found that BPD infants with typical imaging findings were significantly more likely to require HOT at discharge (74.1 vs. 46.6%). The rates of HOT were influenced by age at discharge and discharge policies (40, 41) and varied widely among hospitals. The higher overall incidence of HOT at discharge in this study may be due to the heavy financial burden on the family and the desire for an earlier discharge to facilitate infant bonding at home. This may be why there was no significant difference in hospital stay or duration of oxygen supplementation between the two groups. The target oxygen saturation of HOT is higher than the oxygen-dependent standard, so the ratio of HOT in our study was higher than the ratio of moderate to severe BPD (42).

Our study indicated that typical imaging findings were associated with a higher rate of wheezing disorders (12.2 vs. 5.9%). However, no significant difference was noted between the groups (p = 0.285). If the follow-up period was longer, it would not be surprising to find a positive conclusion. For example, Arai et al. demonstrated that bubbly/cystic appearance on CXR after 28 days in infants with BPD was a potential risk factor for wheezing disorders at 3 years of age (43). Other studies showed that older children and young adult survivors of BPD had an increased risk of asthma (44–46).

Previous studies (47, 48) have shown that infants with BPD were more than twice as likely to be hospitalized during the first year of life as those without BPD. Our data demonstrated that the presence of typical imaging findings did not significantly increase the rate of rehospitalization or the frequency of clinical visits during the first two years after birth. Compared with infants without typical imaging findings, BPD infants with typical imaging findings did not have a higher proportion of underweight or stunting. This may be reflective of changes and improvements in clinical characteristics, postnatal treatment and admission criteria, which allows determination of and confidence in the clinical treatment of patients with typical imaging findings.

The exact etiologies of fibrotic and cystic changes are unclear and are likely to be associated with intrauterine inflammation. Intrauterine infection and CAM are believed to induce fetal lung inflammation, which could interfere with lung development, resulting in BPD. A meta-analysis confirmed that among preterm infants, exposure to CMA was associated with an increased risk of developing BPD (49). Hirata et al. (22) also reported an association between a cystic/bubbly pattern and intrauterine inflammation. The correlation of typical imaging findings with CAM is clouded by the complexity of fetal exposure to inflammation. Additional studies are needed to show the relationship between typical chest imaging findings and CAM.

Some limitations of the study are as follows: (1) this is a retrospective study, which inevitably leads to loss of follow-up and clinical data. A standardized follow-up program has not been perfected in our unit, especially 3 years ago or even earlier. Poor follow-up compliance was related to different geographical origins of enrolled infants and parents' reluctance to travel a long distance. We routinely asked about the condition of the infant after being discharged from the hospital by phone. Some information was not easy to obtain, there was a certain probability of bias. Nevertheless, we tried our best to verify relevant information and only presented reliable data. (2) This is a single-center study of Chinese premature infants. Our unit did not perform the oxygen reduction test, causing some infants to receive a longer time of supplemental oxygen (50). There are some differences in the incidence and severity of some complications from other international studies (51). (3) We used propensity score matching to balance some of the known confounding variables between the groups, making the outcomes more comparable and the conclusion more reliable and stable. However, it is inevitable to pay the cost of missing samples. (4) In addition, we only evaluated the role of typical imaging findings in preterm infants with BPD. Perhaps some infants with similar chest imaging findings would not even have BPD at all. In the later stage, we could work with the radiology department to conduct a larger population with a wider gestational age study.

Despite these limitations, this study provides a new perspective for understanding chest imaging in BPD. Typical chest imaging findings combined with medical history and clinical characteristics can provide powerful help for early diagnosis and treatment.

This study provides important and fresh information that helps explain the potential harms that might be expected when typical imaging findings are noted in infants with BPD. In summary, our results demonstrated that typical imaging findings were significantly associated with the severity of BPD and a higher rate of mortality in this retrospective study. BPD infants with typical imaging findings also accounted for a higher proportion of infants requiring HOT at discharge.

To identify the influence of chest radiographic findings on long-term respiratory and neurodevelopmental outcomes, further prospective randomized controlled studies with larger samples and strict follow-up are needed.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by the Institutional Review Board of Children's Hospital of Chongqing Medical University (2019-209). Written informed consent from the participants' legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements.

QR participated in the design of the study, collected clinical data, conducted statistical analyses, provided interpretation of data, drafted the initial manuscript, and reviewed the final manuscript as submitted. JW conceptualized and designed the study, provided interpretation of data, and revised the article for important intellectual content. YS conceptualized and supervised the design and execution of the study, reviewed analyses, and critically reviewed and revised the manuscript for important intellectual content. All authors contributed to the article and approved the final manuscript as submitted.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.