6573 Validation of NICHD BPD calculator in prediction of BPD in a single neonatal unit, UK- Retrospective study

Objectives Background: Predicting which baby is likely to develop Bronchopulmonary dysplasia (BPD) would help us in management, and counselling parents. Objective: To validate the usefulness of National Institute of Child Health and Human Development (NICHD) BPD calculator in prediction of BPD in our neonatal unit. Methods NICHD calculator: NICHD has developed web based BPD estimator which includes parameters such as gestation age (23–30 wk), birth weight (501–1249 g), sex, ethnicity, postnatal day, ventilator type, non-invasive respiratory support and fraction of inspired oxygen. It provides the probability of outcomes in terms of no-, mild-, moderate-, severe-BPD, and death. Methods: we conducted a single centre retrospective study over a period of 5 months (January to May 2023). We collected all the data using neonatal electronic database on all preterm infants with birth gestational age ≤30 weeks. Accuracy of BPD prediction at three specific time points (day 1, 3, 7) using NICHD calculator was assessed by comparing the calculated risk with observed rates. Receiver operator characteristic (ROC) curves was used to assess the accuracy of the calculator for the outcomes Death or moderate-severe BPD and moderate-severe BPD. We used the BPD definition described by Jensen et al. Mild BPD (Grade 1) for infants receiving low flow oxygen <2L/min, Moderate BPD (Grade 2) for infants receiving low flow oxygen >2L/min or NCPAP/High flow/NIPPV and severe BPD (Grade 3) for infants receiving any invasive positive pressure ventilation. We considered area under the curve (AUC) of 0.5 suggesting no discrimination, 0.7 to 0.8 is considered acceptable, 0.8 to 0.9 is considered excellent, and more than 0.9 is considered outstanding. Results Over the study period we had 38 preterm infants ≤30 weeks. After excluding infants with birth weight >1249grams, we had a total of 29infants for analysis. During the study period we had 5 deaths, 17 infants with 'any BPD', 10 infants with 'grade 1 BPD' and 7 infants had 'grade 2 BPD'. Prevalence of death or mod-severe BPD in our study cohort was 34% (10/29). For calculating ROC curve, we had 23 infants after excluding 6 infants with incomplete data. ROC curve with area under the curve and 95% confidence interval was provided in the figure 1. Conclusion In our small sample size study, NICHD calculator had excellent prediction for the otucomes of Mod-severe BPD. This calculator would be useful in managing high risk preterm infants and needs further validation with larger sample size.