Bronchopulmonary dysplasia (BPD) was initially described by Northway et al. [1] as a chronic lung disease in premature infants, requiring mechanical ventilation and oxygen therapy for neonatal respiratory distress syndrome (NRDS). In recent years, the survival rate for extremely premature infants has significantly increased due to advances in obstetrical and neonatal care, such as antenatal steroid treatment [2], surfactant therapy [3], oxygen saturation target, caffeine, and mechanical ventilation strategies [4]. Meanwhile, promising therapies are being transferred from bench to bedside, including mesenchymal stromal cells (MSCs), insulin-like growth factor 1/binding protein-3 (IGF-1/IGFBP-3), and interleukin 1 receptor (IL-1R) antagonist (anakinra) [5]. Nevertheless, there has not been a decrease in the incidence of BPD [6], with an estimated prevalence ranging from 10.8% to 37.1% among preterm neonates born at 24 to 31⁺⁶ weeks gestational age (GA) and a birth weight (BW) of 1500 g [7].

BPD is characterized by a large and simplified alveolar structure, and a reduced and dysmorphic vascular bed [8]. The pathogenesis of BPD is multifactorial and includes immature lungs, NRDS, barotrauma, volutrauma, oxygen toxicity, sepsis, inflammation and gene specificity. BPD is associated with long-term pulmonary morbidities, such as reduced lung function, increased risk of emphysema, high incidence of wheezing, impaired growth and mental development, as well as cerebral palsy [9, 10]. Despite extensive research, the ability to predict which infants will develop BPD early in life remains challenging. A regression model that improves the accuracy of predicting BPD in preterm infants was recently proposed [11]. This model includes mechanical ventilation lasting $>$7 days, frequent apnea episodes, the time taken to achieve full enteral nutrition, and serum biomarkers such as the level of mediator complex subunit 1 (MED1) and peroxisome proliferator-activated receptor gamma coactivator-1alpha (PGC-1$\alpha$) on day 1 of life (DOL1). Several models have also included tracheal biomarkers, blood tests, and lung ultrasound, but these have yet to be externally validated [12]. Currently, however, there are no reliable indicators for the early diagnosis of BPD, nor for predicting its severity or treatment response. It is therefore important to identify accurate and robust predictors of BPD.

Anemia is common among preterm infants and can result in tissue hypoxia, anaerobic metabolism and the accumulation of lactic acid, as well as giving rise to inflammation [13]. Hemoglobin (Hb) and hematocrit (Hct) levels are significant erythrocyte-related indices that can be derived from routine hematologic parameters. Decreased levels of these markers are characteristic of anemia. Previous studies have reported that lower Hb or Hct levels were closely related to BPD. Duan et al. [14] reported that preterm infants with BPD had lower Hct level than those without BPD on days 1, 7, 14 and 21 after birth, and that early anemia within 14 days after birth was a risk factor for BPD. These researchers also found that Hb $\le$155 g/L within the first 3 days of life was a significant risk factor for BPD [15]. However, the correlation between Hb and Hct levels and BPD at different time points remains unclear. Moreover, studies on the pre-delivery Hb and Hct levels of the mothers of BPD infants have yet to be reported.

Hence, the objective of this study was to evaluate the predictive value for BPD of erythrocyte-related indices, such as Hb and Hct, in preterm infants on days 1, 3, 14 and 28 of life (DOL1, DOL3, DOL14 and DOL28, respectively).

Although BPD is the most common clinical issue of prematurity, accurately predicting this condition at an early stage remains challenging. Clinical prediction models, echocardiogram measurements, lung function tests, epigenetic factors and various biomarkers have all been suggested as potential early indicators of BPD [23, 24, 25], but have yet to be consistently validated. Anemia of prematurity (AOP) is frequently observed in preterm infants, especially in those born at a GA $\le$32 weeks [26]. Previous studies have indicated a link between AOP and complications in preterm infants. For instance, Patel et al. [27] demonstrated that severe anemia (Hb $\le$80 g/L) was associated with a heightened risk of NEC in preterm infants with a BW 1500 g. Additionally, the presence of anemia at birth has been significantly associated with conditions such as hsPDA, severe intra-periventricular hemorrhage (IPVH), and an increased rate of mortality [21, 28, 29]. Duan et al. [14] found that low Hct level were associated with BPD. A retrospective analysis of 147 infants with GA $\le$32 weeks revealed that those with BPD had lower Hb level during the first 3 days of life [15].

Maternal anemia during pregnancy is a known risk factor for anemia in extremely low birth weight (ELBW) infants (BW 1000 g) and extreme preterm infants. However, previous study has not emphasized the connection between maternal anemia and BPD [30]. In the present study, no significant correlations were found between maternal Hb and Hct levels and preterm infants, regardless of whether they developed BPD. We also compared Hb and Hct levels at various time points between preterm infants with or without BPD. BPD patients had lower Hb and Hct levels on DOL1, DOL3 and DOL14, but higher levels on DOL28, possibly due to their earlier PRBC transfusions (Table 1). Our findings suggest that DOL3-Hb $\le$150 g/L could be an early predictor of BPD. Furthermore, the area under the curve (AUC) for Hb on DOL1, DOL3 and DOL14 was greater than that for Hct, indicating it has superior predictive value within the first 14 days of life. Combination of the Hb and Hct levels on DOL3 did not improve the predictive value for BPD. Multivariate regression analysis confirmed that DOL3-Hb $\le$150 g/L was significantly associated with BPD (adjusted OR = 3.222, p = 0.002), highlighting the importance of monitoring Hb level in relation to BPD development. The potential mechanisms linking early anemia to BPD may include a reduction in the blood’s oxygen-carrying capacity, which can lead to increased anaerobic metabolism and elevated lactic acid levels. Additionally, anemia may limit the oxygen supply to the brain’s respiratory center, resulting in respiratory issues such as tachypnea, dyspnea, and apnea, which could prolong the need for supplemental oxygen or mechanical ventilation. Moreover, as preterm infants transition from the fetal environment to early life, their lungs must adapt to new conditions, and Hb plays a crucial role in this adaptation process. Lower Hb level may hinder normal lung function development, potentially contributing to the onset of BPD [31, 32, 33, 34].

AOP is influenced by various factors, such as rapid postnatal growth, blood loss due to medical procedures, and deficiencies in essential micronutrients necessary for red blood cell production [35]. Several clinical practices have been implemented to reduce the severity of AOP. One such practice, delayed cord clamping (DCC), allows blood to transfer from the placenta to the newborn, thereby increasing Hb level and decreasing the need for PRBC transfusions [36]. However, a large randomized trial found that DCC (lasting at least 60 s) did not impact the incidence of BPD in infants born at GA $\le$30 weeks [37], suggesting the effectiveness of DCC in reducing AOP should be reassessed [38]. Phlebotomy losses are significant contributors to AOP, with ELBW infants losing about 11–22 mL/kg of blood weekly during the first 6 weeks of life [39]. The neonatal total circulating volume is approximately 80 mL/kg. Therefore, it is crucial to expand the range of analytes that can be measured using microsample point-of-care devices, to utilize placental blood for routine tests, and to implement neonatal sampling tubes to minimize blood loss.

Typically, the Hb level falls significantly during the first week after birth, leading to AOP [40]. PRBC transfusions are a key intervention for addressing AOP, with up to 40% of very low birth weight (VLBW) infants [41] and 90% of ELBW infants receiving PRBC transfusions during their hospital stay [42]. The Effects of Transfusion Thresholds on Neurocognitive Outcomes of Extremely Low-Birth-Weight Infants (ETTNO) trial and Transfusion of Prematures (TOP) trial have shown that receiving relatively low levels of Hb before transfusion, did not have any negative effect on the incidence of BPD [43, 44]. However, numerous studies have reported that PRBC transfusions were an independent risk factor for BPD in preterm infants. Lee et al. [45] found that frequent PRBC transfusions in the first week of life and a higher total volume of PRBC transfusions were associated with an increased risk of BPD. We also found that PRBC transfusions were more frequent in patients with BPD compared to those without, suggesting it was a significant risk factor. Following PRBC transfusion, erythrocytes release heme, which is subsequently broken down and raises the serum iron level. Free radicals generated by the excess iron can lead to tissue damage. Since premature infants are particularly susceptible to oxidative damage [46], this oxidative stress and the resulting inflammatory cascade may provide an explanation for the development of BPD [47].

Recombinant human erythropoietin (rhEPO) has the ability to promote erythropoiesis and reduce the need for PRBC transfusions in AOP cases [48]. Bui et al. [49] reported that administration of rhEPO (250–300 U/kg) three times a week via intravenous or subcutaneous injection reduced the occurrence of BPD. Research on animals has demonstrated that erythropoietin (EPO) can stimulate angiogenesis [50], enhance alveolar development, reduce lung fibrosis during oxygen exposure, and inhibit transforming growth factor-$\beta$1 (TGF-$\beta$1) signaling pathways by mobilizing endothelial progenitor cells [51]. These findings support the use of EPO for extremely preterm infants who are at high risk for BPD.

The present work has several limitations. First, this was a retrospective study based on data from a single center, and lacks external validation. Larger prospective studies are required to better understand the role and clinical significance of Hb level in the onset and progression of BPD. Secondly, there may be a selection bias since we excluded some patients who either died or were not treated before being diagnosed with BPD. These preterm infants might also be at a higher risk for BPD due to intubation. Thirdly, Villar et al. [52] proposed that preterm birth can no longer be defined by GA alone since this approach fails to provide any pathophysiologic insights or assessment of specific risks. Premature infants are defined by GA ˂37 weeks in our research without considering their phenotypes. In the future study, we will incorporate new theories to define preterm birth and assess the outcomes of different phenotypes on BPD. Despite these limitations, our study is highly relevant to clinical practice and provides new insights that could assist in the management and prediction of outcomes for preterm infants.

In summary, this study demonstrated that patients with BPD had lower Hb and Hct levels up to DOL14. DOL3-Hb $\le$150 g/L may be an early indicator for BPD. Additionally, we identified a link between the number of PRBC transfusions and a higher occurrence of BPD in premature infants. It is important to develop appropriate transfusion protocols and to consider treatments such as rhEPO to help prevent AOP.

BPD, bronchopulmonary dysplasia; MSCs, mesenchymal stromal cells; IL-1R, interleukin 1 receptor; IGF-1/IGFBP-3, insulin-like growth factor 1/binding protein-3; Hb, hemoglobin; Hct, hematocrit; DOL, day of life; GA, gestational age; BW, birth weight; MED1, mediator complex subunit 1; PGC-1$\alpha$, peroxisome proliferator-activated receptor gamma coactivator-1alpha; hsPDA, hemodynamically significant patent ductus arteriosus; IVH, intraventricular hemorrhage; ROP, retinopathy of prematurity; PRBC, packed red blood cell; DOL3-Hb, Hb level on the third day of life; PMA, postmenstrual age; SGA, small for gestational age; AUC, area under the curve; ROC, receiver operating characteristic; IPVH, intra-periventricular hemorrhage; NEC, necrotizing enterocolitis; AOP, anemia of prematurity; VLBW, very low birth weight; DCC, delayed cord clamping; ELBW, extremely low birth weight; rhEPO, recombinant human erythropoietin; EPO, erythropoietin; TGF-$\beta$1, transforming growth factor-$\beta$1; SD, standard deviation; ETTNO, Effects of Transfusion Thresholds on Neurocognitive Outcomes of Extremely Low-Birth-Weight Infants; TOP, Transfusion of Prematures.

All data reported in this paper will also be shared by the corresponding author upon request.

YS and RZ designed the study. CC and SW analyzed the data and performed the statistical analysis. RZ and YL collected data. CL designed of the work. QW interpreted of data of the work. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The study was approved by the Ethics Committee of the First Affiliated Hospital of Wenzhou Medical University (No.KY2021-R083) and followed the Helsinki Declaration. Written informed consent was obtained from parents when patients were admitted to hospital.

We would like to express our gratitude to all those who helped us during the writing of this manuscript. Thanks to all the peer reviewers for their opinions and suggestions.

All phases of this study were supported by Wenzhou Science and Technology Bureau, China (Y20210274); Zhejiang Medical Association clinical research fund project, China (2024ZYC-B51).

The authors declare no conflict of interest.