A nomogram to predict bronchopulmonary dysplasia in very preterm and/or very low birth weight infants with early-onset sepsis.

BACKGROUND

Early-onset sepsis (EOS) may increase the risk of bronchopulmonary dysplasia (BPD) in very preterm and/or very low birth weight infants, but prediction tools tailored to this population are limited. We aimed to develop and validate a hospitalization-based model for predicting BPD in infants with EOS.

METHODS

This multicenter retrospective study included infants with gestational age < 32 weeks and/or birth weight < 1500 g diagnosed with EOS at three tertiary neonatal intensive care units between January 2023 and December 2024. Infants from Qilu Hospital of Shandong University and Beijing Friendship Hospital formed the development cohort (n = 255), and infants from Cangzhou Central Hospital served as the external validation cohort (n = 85). BPD was defined by the requirement for respiratory support at 36 + 0/7 weeks’ postmenstrual age. Predictors were selected using LASSO regression and entered into multivariable logistic regression to construct a nomogram. Model performance was assessed by discrimination, calibration, and Brier score, with internal validation using 1,000 bootstrap resamples.

RESULTS

Among 340 infants, five predictors were retained in the final model: gestational age, surfactant administration, purulent meningitis, late-onset sepsis, and assisted reproductive technology. In the development cohort, the model showed good discrimination (C-index 0.785, 95% CI 0.727–0.842) and overall accuracy (Brier score 0.178). In the external validation cohort, the model demonstrated acceptable discrimination (AUC 0.772, 95% CI 0.667–0.878) and good overall accuracy (Brier score 0.191).

CONCLUSIONS

We developed and validated a pragmatic hospitalization-based model for predicting BPD in very preterm and/or very low birth weight infants with EOS. This EOS-specific tool may support in-hospital risk stratification and targeted prevention during the NICU course. However, further validation in larger and more diverse populations is warranted before routine clinical implementation.