Bronchopulmonary dysplasia (BPD) is a chronic lung disease primarily caused by inflammation and lung injury resulting from mechanical ventilation and supplemental oxygen therapy, which can disrupt alveolar and vascular development [1,2]. Although interventions such as antenatal glucocorticoids, surfactant replacement therapy, and lung-protective ventilation strategies have improved survival rates, they have no significantly reduced BPD among survivors, particularly in extremely preterm (EP) infants [3-5]. In Korea, the overall incidence of moderate-to-severe BPD in very low birth weight infants was approximately 30% [6], with no significant change from 2013 to 2020 [7]. The incidence increases as gestational age (GA) decreases, reaching 50%, 72%, and 87% in infants born at 26−27 weeks, 24−25 weeks, and 22−23 weeks of gestation, respectively [7]. BPD is associated with significant long-term morbidities, including persistent respiratory impairments, increased susceptibility to pulmonary infections, a higher rate of re-hospitalization for respiratory issues, and adverse neurodevelopmental outcomes such as cerebral palsy (CP) and cognitive delays [8-11]. A recent study highlighted the significant effect of BPD on pulmonary function and structure, which has important implications for the treatment and management of preterm infants [12]. As a result, numerous strategies have been studied to prevent and treat BPD, aiming to minimize lung injury, promote lung development, and improve long-term outcomes.

Postnatal systemic corticosteroids, which help reduce lung inflammation, have been widely studied for their potential to prevent and treat BPD. A recent meta-analysis showed that early systemic corticosteroids treatment (initiated within the first 6 days of life) reduced the risk of BPD and combined outcome of mortality or BPD at 36 weeks’ postmenstrual age (PMA) [13]. However, it increased the risks of CP, and the combined outcome of mortality or CP, which these risks being more pronounced with dexamethasone [13]. Due to these adverse effects, multiple national and international organizations recommend against using dexamethasone within the first week after birth to prevent BPD in EP infants [14-16]. Regarding late treatment (delayed until after the first week of life) with systemic corticosteroids, a 2021 Cochrane systematic review demonstrated that only dexamethasone, not hydrocortisone, reduced the risk of BPD and the composite outcome of death or BPD at 36 weeks’ PMA [17]. This meta-analysis reported that dexamethasone did not increase the risk of CP or neurodevelopmental impairment; however, it was noted that the individual studies were underpowered to assess long-term effects [17]. Recent studies have shown that the effect of postnatal systemic corticosteroids, particularly dexamethasone, on survival free of CP varied depending on the risk of BPD [18,19]. Specifically, while dexamethasone had harmful effects in infants at a low risk of BPD, it improved survival free of CP in those at a high risk of BPD [19]. As a result, there is growing consensus that the targeted administration of postnatal corticosteroids should be considered for infants at high risk of severe BPD. However, uncertainty remains regarding the optimal criteria for initiating postnatal corticosteroid treatment.

The respiratory severity score (RSS) is a simple, noninvasive tool for assessing respiratory failure. It can be used continuously without the need for blood sampling and is calculated as the product of mean airway pressure (MAP) and the fraction of inspired oxygen (FiO₂). RSS has been shown to closely correlate with oxygenation index in newborns receiving mechanical ventilation [20]. Furthermore, it has become a valuable tool for predicting BPD severity by quantifying the extent of respiratory support required. A retrospective study conducted by our unit from 2010 to 2014 found strong correlations between RSS on postnatal day (PND) 14, 21, and 28 and the risk of death or severe BPD in EP infants [21]. Recent studies have reported consistent findings [22,23]. Given its potential to stratify risk for BPD, our unit developed and implemented a postnatal systemic corticosteroid protocol in 2015, using RSS as a key criterion for initiating treatment.

This study evaluates the impact of our RSS-guided postnatal systemic corticosteroid protocol on severe BPD and neurodevelopmental impairments over 2 periods: before and after its implementation, in EP infants receiving mechanical ventilation.

During the 2 distinct study periods, a total of 567 neonates born at <GA 28 weeks were admitted. After excluding infants who died within first 7 days (n=36), were not intubated at PND 14 (n=314), had incomplete data (n=7), or had major congenital anomalies (n=2), a total of 208 patients were enrolled. Of these, 108 infants (51.9%) were in phase I, while 100 infants (48.7%) were in phase II.

After the implementation of the RSS-guided postnatal systemic corticosteroid protocol, a greater proportion of infants received steroids at an earlier time point. Among all infants, the proportion with Jensen grade 0 lung status increased in phase II, a trend particularly pronounced among those who did not receive steroids. Among infants receiving steroids, the severity of BPD appeared to improve in phase II compared to phase I. Specifically, the proportion of infants with Jensen grade 3 BPD decreased with borderline statistical significance, whereas the proportion of infants with Jensen grade 1 BPD increased significantly in phase II. There was no significant difference in the proportion of death or any NDI between the 2 phases.

This protocol considered several factors when determining the timing of steroid administration, the target population, as well as the type and dosage of the steroid. First, based on existing studies reporting that early dexamethasone administration within 7 days of life is associated with poor neurological outcomes [13], steroid administration within the postnatal 7 days was avoided in this protocol. Second, before deciding on steroid administration, echocardiography was performed on infants with an RSS score of 4 or higher, or those with increasing ventilator settings, to check for hemodynamically significant PDA or pulmonary hypertension. Existing studies have shown that the presence of a moderate/large PDA with a left-to-right shunt persisting for 7–14 days increased the risk of BPD [28,29], and prolonged exposure to PDA raised the risk of BPD-associated pulmonary hypertension [30]. Therefore, in cases with hemodynamically significant PDA, the treatment of PDA was prioritized before initiating steroid therapy. Early pulmonary hypertension, detected between 7 and 14 days of life, is also known to be associated with an increased risk of moderate/severe BPD at PMA 36 week [24,31]. Severe BPD encompasses various clinical phenotypes, including parenchymal disease and abnormal pulmonary vascular development. Early disruption of pulmonary angiogenesis impairs alveolarization and causes sustained abnormalities in lung structure, leading to BPD and pulmonary hypertension [32]. Given that steroids reduce lung parenchymal inflammation, their effectiveness may be diminished in cases of BPD accompanied by early signs of pulmonary vascular disease. Third, the decision to use dexamethasone as the sole postnatal steroid in this protocol was based on evidence demonstrating its superior efficacy in improving respiratory outcomes, particularly in facilitating extubation and reducing BPD, compared to hydrocortisone [17,33]. The dosage was minimized, considering its potential impact on neurological developmental outcomes. Fourth, this protocol determined steroid use based on RSS levels. RSS is a simple, noninvasive tool for assessing lung disease severity in preterm infants. Multiple studies have demonstrated its efficacy in predicting severe BPD or death in EP infants [21,22,34]. According to the study published by Jung et al. [21], conducted on patients from this unit, the RSS cutoff values for predicting severe BPD or death were 3.0 on PND 14, 3.6 on PND 21, and 3.24 on PND 28. Based on these results, this protocol set the RSS cutoff for initiating steroid administration at 4.

The most significant result of implementing this protocol was the increase in Jensen grade 0 in Phase II. This result was primarily driven by a significant increase in Jensen grade 0 among the infants who did not receive steroids. These findings suggest that the protocol may have effectively identified infants at lower risk of BPD who did not require steroid treatment. According to a recent meta-analysis on the effect of systemic dexamethasone on survival free of CP, as the risk of BPD increased, the risk difference in survival free of CP between the dexamethasone and control groups also increased [19]. Notably, in groups with a BPD risk of less than 30%, dexamethasone had a detrimental effect. Therefore, it is crucial to identify infants at low risk of BPD and avoid unnecessary steroid use. From this perspective, this protocol may have clinical utility in reducing unnecessary steroid use.

Among the infants who received steroids, the increase in Jensen grade 3 BPD and the decrease in grade 1 BPD suggest a shift toward lower overall BPD severity in phase II. This may be associated with the timing of steroid administration. In phase II, steroids were initiated earlier compared to phase I (median PMA: 31.3 weeks vs. 29.0 weeks, median PND: 34.5 days vs, 29.0 days). Additionally, the timing of administration was more consistent in phase II, with a narrower interquartile range of PMA (29.3–35.3 weeks vs. 27.9–30.7 weeks) and PND (22.8–66.0 days vs. 22.5–36.0 days) compared to phase I. Moreover, approximately 75% of infants received steroids within the first 5 weeks of life in phase II. This finding that earlier administration of steroids may have a positive impact on BPD outcomes, is consistent with other existing studies [35,36]. Kwok et al. [36] found that when dexamethasone was administered at 2–3 weeks of chronological age, it was associated with lower rates of severe BPD and/or death, earlier extubation, and decreased respiratory support at discharge compared to administration after 5 weeks. Cuna et al. [35] concluded that moderately late treatment (postnatal 14–28 days) was associated with fewer days on mechanical ventilation and supplemental oxygen compared to delayed treatment (postnatal 29–42 days). Prolonged mechanical ventilation can lead to ventilation-induced lung injury through proinflammatory signaling pathways, exacerbating lung damage and a generalized systemic inflammatory response [37,38], which may be reduce the effectiveness of the anti-inflammatory effect of postnatal steroids when administered later.

Data derived from current evidence, including meta-analyses, are inconclusive on the long-term neurodevelopmental consequences of the late dexamethasone administration [17,32]. In this study, there were no significant differences in death, any NDI, or the composite outcome of both between phase I and phase II, despite a higher rate of steroid use in phase II. Among the infants receiving steroids, the risk of any NDI tended to decrease with borderline statistical significance in phase II; however, this pattern was not observed among infants who did not receive steroids. This may be related to the possibility that implementing this protocol effectively identified infants at high risk of severe BPD, and applying the treatment reduced the risk of severe BPD. However, as this is a retrospective study with a small sample size, it is limited in drawing definitive conclusions about the long-term effects of dexamethasone. Therefore, further studies are needed to confirm these findings.

This study has several limitations. First, its retrospective design may introduce selection bias, and the relatively small sample size limits the generalizability of the findings. In particular, the subgroup analysis presented in Table 4, which involved infants who received corticosteroid treatment (n=19 and n=33), may limit the reliability and robustness of the observed associations. Nevertheless, the study provides meaningful insights by systematically evaluating the impact of an RSS-guided corticosteroid protocol in a high-risk preterm population over an extended period, using a consistent steroid regimen. Second, loss to follow-up, particularly due to referral to nearby regional hospitals, limited the availability of complete long-term neurodevelopmental outcome data. While the follow-up rate of our unit was relatively high, missing data from transferred patients may have introduced bias, potentially underestimating or overestimating the true neurodevelopmental outcomes. Future large-scale, prospective, multicenter studies are warranted to validate the efficacy and safety of RSS-guided corticosteroid administration, to more accurately assess long-term respiratory and neurodevelopmental outcomes, and to minimize biases associated with small sample sizes and incomplete follow-up.

In conclusion, our RSS-guided postnatal systemic corticosteroid protocol effectively identified high-risk BPD infants, enabling targeted and early steroid use. This approach reduced the risk of severe BPD without negatively affecting neurodevelopmental outcomes. By providing a more precise and individualized strategy for postnatal corticosteroid therapy, this study may help optimize BPD management in EP infants.