A Comparison of the 2022 Versus 2011 National Institute of Child Health and Human Development Web-Based Risk Estimator for Bronchopulmonary Dysplasia

Abstract

Objective: To compare the predictive accuracy of the 2022 vs 2011 National Institute of Child Health and Human Development estimator for identifying infants at high risk for bronchopulmonary dysplasia (BPD). Methods: We conducted a single-center retrospective study of infants ≤28 weeks' gestation. Demographic and respiratory support data were used to calculate risk for BPD or death by the BPD estimators. Our outcomes of interest included (1) treatment with systemic steroids for BPD and (2) composite outcome of death or highest severity of BPD at 36 weeks’ postmenstrual age. We used c-statistics and area under receiver operator characteristic curves (AUC) to compare accuracy of the BPD estimators. Results: A total of 159 infants (mean gestational age, 26 weeks; mean birth weight, 837 g) were included. Steroid treatment for BPD occurred in 61 infants. We found that predictive accuracy was similar for both 2022 and 2011 BPD estimators, with c-statistics ranging from 0.77 to 0.89, indicating good-to-excellent accuracy for predicting steroid treatment. Twenty-one infants had death or grade 3 BPD based on definitions used in the 2022 estimator, and 68 infants had death or severe BPD based on definitions used in the 2011 estimator. Overall, we found both BPD estimators had poor AUC for predicting infants at high risk for death or highest severity of BPD, with the exception of the 2022 estimator, which had an acceptable AUC of 0.773 on postnatal day of life 28. Conclusions: The 2022 and 2011 BPD estimators both demonstrate similar good-to-excellent accuracy for identifying infants at high risk for steroid treatment, but poor-to-fair accuracy for predicting death or highest severity of BPD at 36 weeks’ postmenstrual age.

Keywords

• prematurity
• bronchopulmonary dysplasia
• postnatal corticosteroids
• risk