Clinical Report: Identifying Factors Influencing Extended Respiratory Support in Late Preterm Infants
Overview
This study identifies key factors influencing the duration of respiratory support in late preterm infants and develops a predictive model using LASSO-Cox regression. The findings highlight the importance of multiple pregnancy and superoxide dismutase levels as risk factors, while early use of nasal continuous positive airway pressure (NCPAP) appears protective.
Background
Late preterm infants, constituting over 70% of preterm births, frequently require respiratory support due to underdeveloped respiratory systems. Prolonged respiratory support can lead to serious complications, making accurate prediction of weaning timing essential for improving clinical outcomes. Traditional methods often fail to account for complex interactions among multiple predictors, necessitating advanced modeling techniques.
Multiple pregnancy is a significant risk factor for prolonged respiratory support (HR = 1.289).
Higher levels of superoxide dismutase (SOD) are associated with increased risk of extended support (HR = 1.014).
The albumin-to-globulin ratio (A/G) is a potential risk factor (HR = 1.130).
Early use of nasal continuous positive airway pressure (NCPAP) reduces the duration of respiratory support (HR = 0.703).
Lymphocyte percentage (LYM_PC) and lactate dehydrogenase (LDH) levels are protective factors against prolonged support.
The predictive model achieved a C-index of 0.677, indicating moderate discrimination.
Clinical Implications
Clinicians should consider multiple pregnancy and elevated SOD levels as indicators of potential prolonged respiratory support in late preterm infants. Implementing early NCPAP can be beneficial in reducing the duration of respiratory support, thereby improving patient outcomes.
Conclusion
The LASSO-Cox nomogram provides a valuable tool for predicting respiratory support duration in late preterm infants, emphasizing the need for individualized assessment based on identified risk factors.
