Early Predictors of Bronchopulmonary Dysplasia in Extremely Preterm Infants

Overview

Bronchopulmonary dysplasia (BPD) remains one of the most common complications in extremely preterm infants despite advances in neonatal intensive care. Early identification of infants at high risk for BPD could facilitate individualized treatment strategies and improve long-term respiratory outcomes. However, current prediction models rely primarily on conventional clinical variables and have limited predictive accuracy. This prospective observational study aims to evaluate the predictive value of novel physiological, imaging, and biomarker-based indicators for the development of BPD in infants born before 26 weeks of gestation. Participants admitted to the neonatal intensive care units of the Medical University of Vienna will undergo non-invasive assessments during routine clinical care. These include respiratory function monitoring during neonatal transition, lung ultrasound, diaphragmatic ultrasound, targeted neonatal echocardiography, forced oscillation technique, neurally adjusted ventilatory assist-derived diaphragmatic electrical activity, electrical impedance tomography, and proteomic analyses of plasma and tracheal aspirate samples. The association between these novel indicators and respiratory disease severity will be assessed using the Respiratory Severity Score (RSS). Their ability to predict BPD will be evaluated using receiver operating characteristic (ROC) analysis and uni- and multivariable logistic regression models. Predictive performance will be assessed using the C-statistics. The primary objectives are to determine the correlation between the novel indicators and the RSS and to evaluate their diagnostic accuracy for predicting BPD. Secondary objectives include identifying the combination of indicators that provides the best prediction of BPD, evaluating longitudinal changes in respiratory and cardiovascular parameters during the neonatal period, assessing the effects of respiratory interventions and treatments on these indicators, and investigating associations with survival and major neonatal morbidities. A total of 140 extremely preterm infants is planned for inclusion in the study. The results are expected to improve early risk stratification and contribute to the development of individualized strategies for preventing and managing BPD.

Background: Bronchopulmonary dysplasia (BPD) remains one of the most frequent and clinically relevant complications of extreme prematurity. Although advances in perinatal and neonatal care have substantially improved the survival of infants born at the limits of viability, the incidence of BPD has remained relatively unchanged. BPD is associated with prolonged respiratory support, recurrent respiratory infections, pulmonary hypertension, impaired lung function throughout childhood and adulthood, adverse neurodevelopmental outcomes, and increased healthcare utilization. Current prediction of BPD is primarily based on conventional clinical variables, including gestational age, birth weight, duration of respiratory support, oxygen requirement, and the Respiratory Severity Score (RSS), calculated as the product of mean airway pressure (MAP) and the fraction of inspired oxygen (FiO₂). Although these parameters provide valuable information regarding disease severity, they do not directly characterize the underlying pathophysiological mechanisms of lung injury or capture the complex interactions between pulmonary mechanics, ventilation distribution, respiratory muscle function, cardiovascular adaptation, and inflammatory responses. Several novel monitoring technologies have become available in neonatal intensive care over the past decade. These include lung ultrasound, targeted neonatal echocardiography, respiratory function monitoring, electrical impedance tomography, neurally adjusted ventilatory assist, forced oscillation technique, and advanced proteomic analyses. These technologies provide quantitative physiological information that may identify infants at increased risk of developing BPD before irreversible lung injury occurs. However, their predictive value when evaluated in combination has not yet been systematically assessed in a comprehensive prospective cohort. This study aims to investigate whether physiological, imaging, and molecular biomarkers obtained during the neonatal period improve the early prediction of BPD compared with conventional clinical indicators. Study Objectives: The primary objective is to determine the predictive value of novel physiological, imaging, and biomarker-based indicators for the development of bronchopulmonary dysplasia in extremely preterm infants. Specific objectives include: * evaluating the association between novel measurements and the respiratory severity score (RSS); * determining the diagnostic performance of each parameter for predicting BPD; * identifying combinations of predictors that provide optimal discrimination between infants who develop BPD and those who do not; * evaluating longitudinal changes of physiological parameters during the first weeks of life; * investigating the effects of respiratory interventions and routine clinical management on these measurements. Study Design: This is a prospective, single-center, observational clinical study performed at the neonatal intensive care units of the Medical University of Vienna. No experimental intervention will be performed. All patients will receive standard neonatal intensive care according to local protocols. Study-related assessments consist of non-invasive physiological measurements that are performed alongside routine clinical care whenever possible. Study Population: Eligible participants are infants born before 26 completed weeks of gestation who are admitted to the neonatal intensive care units of the Medical University of Vienna. Written informed consent will be obtained from parents or legal guardians before study inclusion according to institutional regulations. The a-priori sample size calculation was based on previously published data on the predictive performance of lung ultrasound (LUS) and the Forced Oscillation Technique (FOT). Assuming a BPD incidence of approximately 30%, a statistical power of 95%, a two-sided significance level of 5%, and an anticipated dropout rate of 30%, the required sample size was estimated to be 140 infants. Study Procedures: Each participant will undergo repeated physiological and imaging assessments during the neonatal period. 1. Respiratory Function Monitoring During neonatal stabilization immediately after birth, respiratory function monitoring will be used to continuously record ventilatory parameters during mask ventilation, CPAP, or mechanical ventilation. Measured variables include: * peak inspiratory pressure, * expiratory tidal volume, * respiratory rate, * leak around the face mask, * airway pressure curves, * ventilation effectiveness. These measurements characterize the quality of respiratory support during neonatal transition and may provide early markers of lung injury. 2. Lung Ultrasound: Serial lung ultrasound examinations will be performed during the first 28 days of life. The Lung Ultrasound Score (LUS) will be calculated using standardized scoring systems reflecting pulmonary aeration. Additionally, diaphragmatic ultrasound will quantify: * diaphragmatic excursion, * diaphragmatic contractility, * respiratory muscle function. These measurements provide information regarding pulmonary aeration and respiratory muscle performance. 3. Targeted Neonatal Echocardiography Targeted neonatal echocardiography (TnEcho) will be performed repeatedly from birth until 36 weeks' postmenstrual age. Cardiovascular assessments include evaluation of: * ductus arteriosus, * pulmonary artery pressure, * ventricular function, * pulmonary vascular resistance, * systemic blood flow, * pulmonary hypertension, * cardiac output. Cardiopulmonary interactions are expected to influence respiratory outcomes and may contribute to BPD development. 4. Forced Oscillation Technique The Forced Oscillation Technique (FOT) will be performed during the first two postnatal weeks to assess respiratory mechanics. Primary measurements include respiratory system reactance and resistance. Additional follow-up assessments will be performed at approximately 3 years and 5 years of age to investigate long-term pulmonary function in survivors. 5. Neurally Adjusted Ventilatory Assist In infants receiving NAVA ventilation, the electrical activity of the diaphragm (EAdi) will be continuously recorded. Variables include: * peak EAdi, * minimum EAdi, * neural respiratory drive, * breathing effort. These measurements quantify respiratory muscle activation and patient-ventilator interaction. 6. Electrical Impedance Tomography Electrical impedance tomography (EIT) will be performed repeatedly during the first 14 postnatal days. EIT provides continuous bedside imaging of regional lung ventilation without ionizing radiation. Derived parameters include: * center of ventilation, * ventilation distribution, * global inhomogeneity index, * overdistension, * atelectasis/collapse, * regional ventilation delay, * end-expiratory lung impedance changes. Measurements will also be obtained during clinically indicated PEEP or continuous distending pressure (CDP) optimization procedures. 7. Proteomic Analysis Blood plasma and tracheal aspirate samples will be collected at predefined intervals during the first 28 days of life whenever clinically indicated samples are obtained. Proteomic analyses will investigate inflammatory mediators, growth factors, extracellular matrix proteins, and additional molecular biomarkers associated with lung injury and repair. The aim is to identify molecular signatures associated with subsequent development of BPD. 8. Clinical Data Collection Clinical variables routinely collected during hospitalization include: * gestational age, * birth weight, * sex, * antenatal corticosteroid exposure, * Apgar scores, * surfactant administration, * respiratory support, * ventilator settings, * oxygen supplementation, * respiratory severity score, * duration of invasive ventilation, * duration of non-invasive respiratory support, * postnatal corticosteroid therapy, * pulmonary hypertension, * patent ductus arteriosus, * late-onset sepsis, * necrotizing enterocolitis, * intraventricular hemorrhage, * retinopathy of prematurity, * survival, * duration of hospitalization. Statistical Analysis: Continuous variables will be summarized using appropriate descriptive statistics according to their distribution. Categorical variables will be presented as frequencies and percentages. Novel physiological indicators will first be correlated with the Respiratory Severity Score (RSS), calculated as: RSS = Mean Airway Pressure × Fraction of Inspired Oxygen (MAP × FiO₂) Correlation analyses will use Pearson or Spearman coefficients depending on data distribution. The predictive ability of each novel indicator for BPD will be evaluated using receiver operating characteristic (ROC) analysis. For each parameter, the following will be calculated: * area under the ROC curve (AUC), * optimal cut-off values, * sensitivity, * specificity, * positive predictive value, * negative predictive value, * likelihood ratios. Logistic regression analyses will be performed to investigate associations between candidate predictors and BPD. Initially, univariable logistic regression models will be fitted for each predictor separately. Subsequently, multivariable logistic regression models will evaluate combinations of physiological variables while adjusting for important clinical covariates such as gestational age, birth weight, sex, and antenatal corticosteroid exposure. Model performance will be assessed using: * C-statistic (AUC), * calibration, * odds ratios with 95% confidence intervals, * likelihood ratio tests, * internal validation where appropriate. Mutual correlations among novel biomarkers will also be investigated to determine complementary or overlapping information provided by the different technologies. Primary Outcomes: 1. Correlation between each indicator and the Respiratory Severity Score. 2. Diagnostic accuracy of each predictor for bronchopulmonary dysplasia as assessed by receiver operating characteristic analysis. Secondary Outcomes: Secondary analyses include: 1. prediction of BPD using individual physiological parameters; 2. comparison of multivariable prediction models; 3. identification of the strongest independent predictors of BPD; 4. comparison of physiological measurements in intubated and non-intubated infants; 5. longitudinal changes in lung ultrasound, diaphragmatic ultrasound, targeted echocardiography, forced oscillation measurements, electrical diaphragmatic activity, electrical impedance tomography, and proteomic biomarkers during the first four postnatal weeks; 6. physiological changes observed during routine PEEP or continuous distending pressure optimization procedures; 7. effects of respiratory interventions, surfactant administration, corticosteroid therapy, and other treatments on physiological measurements; 8. associations between physiological markers and major neonatal morbidities; 9. survival until hospital discharge; 10. respiratory outcomes during early childhood; 11. longitudinal pulmonary function assessed by Forced Oscillation Technique at 3 and 5 years of age. Expected Significance: This study will comprehensively evaluate several innovative physiological monitoring techniques within a single cohort of extremely preterm infants. By integrating respiratory mechanics, lung imaging, cardiovascular assessment, respiratory muscle activity, ventilation distribution, and molecular biomarkers, the study aims to identify early predictors of bronchopulmonary dysplasia that outperform currently available clinical indicators. Improved early risk stratification may facilitate individualized respiratory management, optimize therapeutic decision-making, improve patient selection for future interventional studies, and ultimately contribute to reducing the burden of bronchopulmonary dysplasia in extremely preterm infants.