This comprehensive retrospective study of risk factors for BPD in very preterm infants provides several important insights into the differential pathophysiology of mild versus moderate-to-severe disease. Our findings confirm the hypothesis that while some risk factors are common to all forms of BPD, others are specifically associated with more severe disease. This stratified approach to risk factor analysis represents an important advancement over previous studies that treated BPD as a binary outcome.

Although mild BPD may have minimal long-term clinical impact compared to more severe forms, analyzing it as a distinct group helps delineate the gradient of disease severity and identify factors driving progression from mild to severe forms. Understanding this continuum is essential for developing targeted interventions that may prevent progression and for better characterizing the pathophysiological spectrum of this heterogeneous condition.

One of the most striking findings of our study is the differential association of inflammatory processes with BPD severity. Histological chorioamnionitis, which indicates an intrauterine inflammatory environment, was independently associated with moderate-to-severe BPD but not with mild disease after adjusting for confounders. This suggests that prenatal inflammation may trigger a cascade of events that leads to more severe lung injury and aberrant repair mechanisms. The association was particularly strong in infants born at 29–31 weeks gestation, suggesting that the timing of the inflammatory insult relative to the stage of lung development may be crucial in determining its impact on BPD severity. These findings are consistent with experimental evidence showing that inflammation can disrupt alveolar and vascular development, key pathological features of severe BPD [13].

Similarly, late-onset sepsis emerged as an independent risk factor for moderate-to-severe BPD but not for mild disease. The stronger association in the most immature infants (gestational age < 28 weeks) highlights the vulnerability of this group to postnatal inflammatory insults. Sepsis may contribute to BPD through various mechanisms, including direct lung injury from inflammatory mediators, increased oxygen requirements, and the need for prolonged mechanical ventilation [14]. Our results emphasize the importance of infection prevention strategies, such as strict hand hygiene, central line care bundles, and antimicrobial stewardship, particularly for the most immature infants.

Pulmonary hypertension showed a remarkably strong association with moderate-to-severe BPD, with much weaker association with mild disease. This finding aligns with the concept that pulmonary vascular disease is a central component of severe BPD, often referred to as the “vascular hypothesis” of BPD [15]. Whether pulmonary hypertension is a cause or consequence of severe BPD remains debated, but our findings suggest that strategies targeting pulmonary vascular development and function might be particularly beneficial for preventing progression to severe disease. Early screening for pulmonary hypertension in high-risk infants and targeted therapies for those with early signs of pulmonary vascular disease warrant further investigation.

The duration of mechanical ventilation showed a dose-dependent relationship with BPD severity, with stronger association for moderate-to-severe than for mild disease. It is important to note that prolonged mechanical ventilation may represent both a cause and a marker of disease severity, reflecting a complex bidirectional relationship. While mechanical ventilation is often necessary for the survival of extremely preterm infants and may contribute to lung injury through barotrauma and volutrauma, the need for prolonged ventilation may also indicate more severe underlying lung pathology. Our findings reinforce the importance of lung-protective ventilation strategies and early extubation whenever possible [16]. Non-invasive respiratory support techniques, such as nasal continuous positive airway pressure and high-flow nasal cannula, may help reduce ventilator-induced lung injury while providing adequate respiratory support. Future research should focus on developing ventilation protocols specifically designed to minimize progression from mild to more severe forms of BPD.

Interestingly, our study found a protective effect of maternal preeclampsia against moderate-to-severe BPD, particularly in infants born at 29–31 weeks gestation. This seemingly paradoxical finding has been reported in previous studies and might be explained by several mechanisms [17,18,19]. Preeclampsia is associated with placental insufficiency and intrauterine growth restriction, which may lead to antenatal stress and accelerated lung maturation. Additionally, the anti-angiogenic environment in preeclampsia might paradoxically promote more normalized postnatal pulmonary vascular development in premature infants. The finding that the protective effect was significant only in the more mature infants suggests that the timing of delivery relative to the preeclamptic insult may be important. However, this protective effect should be interpreted cautiously, as preeclampsia is associated with other adverse outcomes, including growth restriction and neurodevelopmental impairment.

Regarding the lack of association between small-for-gestational-age (SGA) status and BPD in our study, which differs from some previous reports, several explanations are possible. First, there may be collinearity between SGA and gestational age or maternal preeclampsia in our cohort, potentially masking the independent effect of growth restriction in multivariable models. Second, our population may have had a predominance of inflammatory rather than growth-restriction phenotypes of preterm birth, which could influence the risk factor profile. Third, differences in the definition and assessment of SGA across studies may contribute to inconsistent findings. Future research with larger sample sizes and more detailed characterization of growth restriction patterns may help clarify this relationship.

Male sex remained an independent risk factor for both mild and moderate-to-severe BPD after adjusting for other factors, with slightly stronger association for more severe disease. The biological basis for this sex disparity is not fully understood but may involve differences in lung development, surfactant production, and inflammatory responses between males and females [20,21,22,23]. This non-modifiable risk factor is important for risk stratification and prognostication but also highlights the need for research into the mechanisms underlying sex differences in BPD susceptibility.

Lower gestational age emerged as the strongest independent risk factor for both mild and moderate-to-severe BPD, emphasizing that lung immaturity remains the fundamental predisposing factor for this disease. The stronger association with moderate-to-severe BPD reflects the particularly vulnerable state of the most immature lungs to various injurious stimuli. While gestational age itself is not modifiable once an infant is born, this finding underscores the importance of preventing extremely preterm birth whenever possible and optimizing antenatal care when preterm delivery is inevitable.

Our prediction models demonstrated good discriminative ability for both mild and moderate-to-severe BPD, with better performance for the latter. These models could potentially be used for early risk stratification, allowing for targeted preventive interventions in high-risk infants. The population attributable risk analysis further highlighted the potential impact of preventive strategies targeting modifiable risk factors, particularly prolonged mechanical ventilation and late-onset sepsis.

The strengths of our study include its relatively large sample size of extremely preterm infants, comprehensive assessment of risk factors, stratification by BPD severity, and robust statistical analysis accounting for potential confounders. However, several limitations should be acknowledged. First, as a retrospective study, it is subject to the inherent limitations of this design, including potential information bias and unmeasured confounding. Second, while we used the widely accepted NICHD/NHLBI criteria for BPD classification, newer physiological definitions, such as the oxygen reduction test, might provide more accurate assessment of disease severity [24]. Third, our study was conducted at a single tertiary center, potentially limiting the generalizability of our findings to centers with different management practices, particularly given that ventilatory management strategies are highly center-dependent. Fourth, we could not assess genetic factors, which likely play an important role in BPD susceptibility and severity [25].

Our findings have several important clinical implications. First, they suggest that prevention strategies may need to be tailored to the risk profile for different severities of BPD. For example, aggressive management of chorioamnionitis and early screening for pulmonary hypertension might be particularly important for preventing progression to severe BPD. Second, the strong association between late-onset sepsis and moderate-to-severe BPD emphasizes the importance of infection prevention in extremely preterm infants. Third, the dose-dependent relationship between mechanical ventilation duration and BPD severity reinforces the need for lung-protective ventilation strategies and early extubation when possible.

Future research should focus on prospectively validating our prediction models in different populations and exploring the mechanisms underlying the differential associations of risk factors with BPD severity. Longitudinal studies incorporating biomarkers of inflammation, oxidative stress, and vascular development could provide further insights into the pathophysiological pathways leading to different BPD phenotypes. Additionally, clinical trials of preventive interventions should consider stratifying outcomes by BPD severity to better assess their impact on the disease spectrum.

In conclusion, our study demonstrates that risk factors for BPD in very preterm infants(< 32 weeks’ gestation) differ by disease severity, suggesting distinct pathophysiological mechanisms for mild versus moderate-to-severe disease. While lower gestational age and male sex are common risk factors for all forms of BPD, inflammatory processes, pulmonary hypertension, and prolonged mechanical ventilation show stronger associations with moderate-to-severe disease. These findings support a more nuanced approach to BPD prevention and management, taking into account the severity-specific risk profiles identified in this study. By targeting modifiable risk factors associated with more severe disease, we may be able to reduce the substantial burden of severe BPD and improve long-term outcomes for very preterm infants.