Persistent patent ductus arteriosus(PDA) in preterm infants is associated with various neonatal morbidities, including bronchopulmonary dysplasia, intraventricular hemorrhage, pulmonary hemorrhage, necrotizing enterocolitis (1). Unlike full-term infants, the incidence of PDA exceeds 50% in preterm infants born at less than 28 weeks’ gestation (2). Closure of the ductus arteriosus is influenced by a variety of physiologic, molecular, structural factor (3). Among these, prostaglandin E2(PGE2) plays a key role as a potent vasodilator. During fetal development, PGE2 levels remain high, but, they rapidly decline after birth due to decreased placental production and increased catabolism in the lungs (4). Pharmacologic closure of PDA primarily targets prostaglandin inhibition. Inflammation is increasingly recognized as a contributing factor in PDA. Chorioamnionitis, which induces a fetal inflammatory response, is associated with an increased risk of PDA (5). Conversely, antenatal steroids with anti-inflammatory effects reduce the risk of PDA (6). Elevated inflammatory biomarkers in the umbilical cord, such as IL-6, IL-8, IL-10, and IL-12, have also been associated with PDA development (7).

PDA represents a major challenge in neonatal care. While small PDA does not usually affect the hemodynamic circulation in preterm infants, larger ones can cause pulmonary overcirculation and decreased systemic perfusion, termed hemodynamically significant PDA(hsPDA). Currently, the terms PDA and hsPDA are often used interchangeably in preterm infants. Recent studies suggest that early extubation in preterm infants improves outcomes by reducing hospital stay, mortality, bronchopulmonary dysplasia, and intraventricular hemorrhage (8, 9), whereas extubation failure—frequently associated with hsPDA—has a markedly adverse impact on prognosis (10). In addition, fluid restriction, which is often employed as a conservative strategy for hsPDA, may impair nutrition, growth, and organ perfusion. Therefore, early identification of hsPDA is crucial to guide timely, individualized treatment strategies and ultimately improve neonatal outcomes. Although echocardiography remains the diagnostic gold standard, its interpretation depends on operator expertise and clinical context, such as the presence of an atrial septal defect or patent foramen ovale and intravascular volume status. Predicting hsPDA development based solely on the clinical presentation at birth remains difficult, highlighting the need for reliable early biomarkers and novel predictive approaches.

CD71+ erythroid cells (CECs) are immature red blood cells, including erythroblasts and reticulocytes, characterized by co-expressing transferrin receptor 1 (CD71) and erythroid lineage markers, glycophorin A (CD235a) in humans (11). While CECs are almost exclusively found in the bone marrow of adults, they are abundant in the peripheral blood of neonates but decrease with age (12, 13). Beyond erythropoiesis, CECs possess immunomodulatory functions that influence both innate and adaptive immune responses (11, 12). The immunosuppressive effects contribute to neonates’ increased susceptibility to infections but also facilitate colonization of commensal microorganisms after birth (11, 14). B-type natriuretic peptide (BNP) and NT-proBNP are well-established cardiac biomarkers that reflect myocardial wall stress in response to volume overload and are used in diagnosing hsPDA. However, BNP levels typically peak at 48–72 hours after birth and require serial measurements for reliable interpretation. In contrast, CECs reflect endothelial activation—an upstream pathophysiological response to perinatal hemodynamic stress and vascular injury. Therefore, investigating CECs measured within the first 12 hours of life as an early biomarker may offer additional value for very early risk stratification of infants who will subsequently develop hsPDA.

An early blood-based biomarker that identifies preterm infants at high risk for hsPDA before echocardiographic confirmation could improve clinical decision-making by enabling risk stratification and more timely, individualized treatment. Given the established link between inflammation and hsPDA and the immunomodulatory role of CECs, we hypothesized that circulating CEC levels in the first day of life are associated with the subsequent development of hsPDA in preterm infants and may serve as a predictive biomarker. In addition, hemodynamic stress and tissue hypoxia associated with hsPDA may stimulate compensatory erythropoiesis, potentially altering circulating CEC populations. Therefore, in this study we aimed to examine whether the abundance of CECs is associated with hsPDA in preterm infants. To address this, we measured CEC levels in a single peripheral blood sample collected within 12 hours of birth and examined how these values varied with gestational age, compared CEC levels between infants with and without hsPDA, and evaluated whether early CEC measurement could serve as a predictive biomarker for hsPDA development.

To our knowledge, no previous animal or human studies have explored the association between CD71+ erythroid cells and hsPDA in preterm infants.

This study demonstrates that CECs with immunomodulatory effects decrease with increasing gestational age in preterm infants. Importantly, higher CEC levels were significantly associated with an increased risk of developing hsPDA. These results suggest that CECs may serve as a novel biomarker for early identification of preterm infants at risk for hsPDA.

Previous studies have reported a significant association between perinatal inflammation and PDA in preterm infants (5, 6). However, our study showed no significant differences in CRP levels or in the incidence of maternal chorioamnionitis and PROM between groups. This suggests that traditional inflammatory indices may not fully detect the risk of hsPDA, further emphasizing the value of novel biomarkers such as CECs. Meanwhile, our results about CECs are consistent with recent studies indicating that CECs can exert complex immunomodulatory effects, including promoting pro-inflammatory cytokine release under certain conditions. Miller et al. (17) demonstrated that when neonatal CD71+ erythroid cells come into direct contact with maternal mononuclear immune cells, they induce the release of pro-inflammatory cytokines while decreasing TGF-β secretion. Additionally, in preterm infants born after preterm labor, depletion of CD71+ erythroid cells from cord blood mononuclear cells resulted in decrease in myeloid cell secreting IL-6, IFN-γ, and IL-4 in the presence of microbial products. Furthermore, preterm labor derived CD71+ erythroid cells were shown to enhance early CD8+ T cell activation. These findings suggest that maternal–fetal immune crosstalk can “prime” fetal CECs toward a more pro−inflammatory phenotype, indicating that CECs may not only reflect a state of immunosuppression but also participate in inflammatory processes that could contribute to the persistence of hsPDA. This supports the concept that maternal–fetal immune crosstalk may prime fetal CECs toward a pro−inflammatory phenotype, which in turn could promote the persistence of hsPDA. Further mechanistic studies are necessary to elucidate this pathway, as it remains unclear whether elevated CEC levels are a cause, a consequence, or simply an epiphenomenon of the inflammatory and hypoxic condition associated with hsPDA.

As CD71+ erythroid cells are immature erythrocytes, CECs levels were inversely correlated with hemoglobin and other hematologic parameters, which generally increased with gestational age. This trend explains the significant differences observed between the hsPDA and control groups in hematologic parameters (Table 2). In multivariable analysis, CECs remained a robust predictor of hsPDA risk, outperforming hemoglobin, RBC, hematocrit, and platelet levels after adjustment for gestational age (Table 3). This suggests that CECs may integrate both hematologic and immunologic alterations linked to hsPDA risk more effectively than individual traditional hematologic indices. Additionally, in a descriptive analysis restricted to infants with hsPDA, higher early CEC levels tended to be observed in those who later demonstrated a larger ductal diameter (≥ 2.5 mm) on echocardiography. Although this trend did not reach statistical significance because of the limited sample size, it raises the possibility that early CEC elevation may be linked to subsequent PDA enlargement, a hypothesis that should be addressed in larger prospective studies. Despite the high post−hoc power (92.3%), the small number of hsPDA events (n=10) and the lower gestational age of infants with hsPDA—given that CEC levels are strongly related to gestation—may still limit the precision and generalizability of the logistic regression estimates, and the results should therefore be interpreted with caution, as a residual risk of type II error cannot be entirely ruled out.

Higher WBC counts were associated with a lower likelihood of hsPDA even after adjusting for gestational age in our cohort. Previous studies have reported inconsistent associations between platelet counts or anemia and hsPDA, and low hemoglobin levels at birth have also been linked to increased mortality in preterm infants (18–21). Because hematologic parameters change with gestational age and are related to morbidity and mortality in preterm infants, larger prospective studies including reticulocyte count are needed to clarify whether these blood cell indices can serve as reliable predictors of hsPDA or related adverse outcomes. Since CECs are immature erythrocytes, their levels tend to increase in cases of anemia, which explains the inverse relationship observed between CECs and hemoglobin. Our study also confirmed that platelet counts negatively correlate with CECs. Although Chen et al. (2014) reported that lower hemoglobin levels increased the risk of hsPDA in VLBW infants, their analysis did not adjust for gestational age (22). In our study, after performing multivariable analysis with adjustment for gestational age, CECs demonstrated a better explanatory power for hsPDA occurrence than hemoglobin. In line with these findings, ROC analysis demonstrated that early CEC measurement has good discriminative ability for predicting hsPDA, with an AUC of 0.84 in the overall cohort and 0.82 in infants ≤28 weeks of gestation. The relatively high specificity, particularly in the very preterm subgroup, suggests that CEC-based stratification could help identify infants at truly high risk for hsPDA and thereby support more targeted use of echocardiography and therapeutic interventions. While this study focused solely on the relationship between CECs and hsPDA, further research is warranted to explore their associations with other morbidities.

Correa-Rocha et al. confirmed that preterm infants exhibit marked leukopenia and lymphopenia (23). Similarly, in our study, WBC counts tended to increase with gestational age, but the correlation was weak. WBC levels are influenced by a wide range of factors; for example, they are naturally higher after vaginal delivery compared to cesarean section (24) and can also increase due to steroid administration or intrauterine stress (25). Conversely, maternal hypertensive disorders and medication use may lead to decreased fetal WBC counts (25, 26). Importantly, even after adjusting for gestational age in the multivariable model, higher WBC counts were significantly associated with a lower likelihood of hsPDA occurrence, which was an unexpected yet interesting finding. The lower hemoglobin, hematocrit, and hematologic parameters observed in infants with hsPDA may contribute to reduced blood viscosity. According to Poiseuille’s law, which relates flow to viscosity, decreased blood viscosity would theoretically increase volumetric flow through the patent ductus arteriosus, promoting hemodynamic significance (27–29). Conversely, the protective association between higher WBC counts and reduced hsPDA risk may reflect the contribution of blood viscosity to ductal hemodynamics, as elevated WBC can increase blood viscosity and reduce ductal shunt volume. These hematologic parameters therefore represent potentially modifiable factors in the hemodynamic pathophysiology of hsPDA. Future studies should explore these subsets in greater depth to better understand which specific immune cell types are associated with hsPDA development.

Our study is the first to present correlation between CEC values and gestational age, and to investigate their association with hsPDA in preterm infants. While echocardiography remains the gold standard for hsPDA diagnosis, its application in very preterm infants is technically challenging and sometimes unreliable due to operator dependency and clinical instability. Early identification of at-risk infants using CECs could enhance neonatal care by informing risk stratification and enabling targeted interventions. BNP and NT proBNP are well validated cardiac biomarkers that correlate with ductal size, shunt magnitude, and the need for hsPDA treatment, and they are already used in clinical practice with established assay platforms and cut off values (30–32). In contrast, CECs reflect endothelial immaturity and activation rather than myocardial wall stress, and in our cohort elevated CEC levels within the first 12 hours after birth preceded the echocardiographic diagnosis of hsPDA, suggesting a role in very early risk stratification before BNP–guided or echocardiographic decision making. Therefore, CECs should not be viewed as a replacement for BNP at this stage but as a complementary biomarker that may help identify high risk infants for targeted echocardiography and subsequent BNP based monitoring, a concept that requires validation in larger prospective studies.

For CECs to be adopted as a biomarker in clinical practice, testing should be feasible within routine neonatal workflows, inexpensive, rapid, and analytically reliable. In this study, CECs were measured using flow cytometry, which offers high analytical specificity but is less practical for universal screening of preterm infants because it requires specialized equipment, trained personnel, and a turnaround time of approximately 24–48 hours. Although flow cytometry reagents add cost, the overall expense is still lower than that of a single echocardiographic examination and becomes more cost-effective when applied at scale. As an instrument−based method, flow cytometry also allows standardization of sample processing, which is expected to yield good reproducibility. For broader clinical implementation, the development of simpler, potentially automated CEC assays would further enhance feasibility. Ongoing work into agents able to modulate CEC expansion further underscores the potential for CEC-directed therapies in the management of hsPDA. By identifying infants with elevated CEC levels, clinicians could potentially tailor pharmacologic or interventional strategies to those at highest risk of hsPDA−related complications. Several studies are currently developing agents to prevent CEC expansion (14), and our research provides a foundational basis suggesting that pharmacological modulation of CECs could serve as a new therapeutic approach for hsPDA in the future.

Several limitations should be acknowledged. This was a single center study conducted in a Korean tertiary NICU with a limited number of hsPDA cases, which constrains generalizability to other racial and ethnic populations and to different clinical practice settings. Because this was an observational study with a small number of hsPDA events, our analyses can only demonstrate associations rather than prove a causal effect of CD71+ erythroid cells on hsPDA outcomes. This study also lacks an external validation cohort, so the predictive performance of CECs may be overestimated. All infants born before 37 weeks’ gestation were analyzed as a single preterm group, which may mask gestational age–specific differences in CEC biology and hsPDA risk. Restricted blood sample volumes precluded cytokine profiling and functional assays, which are needed to directly address the underlying inflammatory mechanisms. Future research using cytokine analysis and measurement of more specific inflammatory markers, including procalcitonin will be essential to substantiate and clarify the immune pathways involved in CEC-mediated ductal patency. CEC levels were measured within the first 12 hours of life, whereas echocardiography for hsPDA was performed later at days 3–7. At birth, despite patent and often large ductal lumens, left-to-right shunting through the ductus arteriosus is minimal because pulmonary vascular resistance remains elevated. Therefore, echocardiographic assessment at birth would not reliably identify hemodynamic significance. By performing echocardiography at days 3–7, we captured the window when ductal hemodynamic effects become clinically relevant and detectable. However, this time difference between CEC measurement and echocardiographic assessment may have influenced our findings and should be considered when interpreting the results. In addition, although CECs and inflammatory markers were sampled at the same time as part of routine NICU blood work, the dynamic nature of postnatal inflammation means that residual timing effects cannot be completely excluded. Finally, we could not fully control perinatal factors that may affect CEC levels, such as the exact timing of cord clamping, the interval between antenatal steroid administration and delivery, and intrauterine hypoxia. Prior studies have shown that delayed cord clamping improves hemoglobin levels and protects against iron-deficiency anemia (33–35). In the present study, none of the infants underwent delayed cord clamping, and the frequencies of antenatal steroid use and hypoxia-related conditions (pregnancy-induced hypertension, placental abruption, meconium staining) were similar between groups. However, these factors may still have affected CEC levels and should be regarded as potential sources of residual confounding. Future multicenter studies with larger cohorts, expanded immunophenotyping, and standardized perinatal practices are needed to validate our findings. In addition, future work should account for the heterogeneity of preterm infants by stratifying them into finer gestational-age and clinical subgroups and by characterizing longitudinal trajectories of CEC levels to refine gestation-specific reference ranges. Applying modern causal-inference approaches, such as target trial emulation, will also be important to determine whether early CEC elevation has a causal impact on hsPDA risk in distinct patient phenotypes.

In summary, CECs decrease with increasing gestational age, and higher CEC levels on the first day of life show a statistically significant association with the subsequent development of hsPDA in preterm infants. These findings suggest that CECs could be used as predictive biomarkers for hsPDA, with potential to improve neonatal outcomes. Further large-scale studies are necessary to confirm these results and to clarify the underlying mechanisms underlying CEC-related risk of hsPDA.