Predicting Risk for Bronchopulmonary Dysplasia: Selection Criteria for Clinical Trials

Abstract

Early identification of neonates in whom bronchopulmonary dysplasia is most likely to develop permits appropriate enrollment into clinical trials testing early intervention therapies for the prevention or treatment of bronchopulmonary dysplasia. Analysis of 160 neonatal intensive care unit survivors to 28 days revealed that supplemental oxygen requirement at 28 days could be predicted by a logistic regression including (1) birth weight, gestational age, 5-minute Apgar score, and peak inspiratory pressure at 12 hours for 12-hour-old neonates and (2) birth weight, gestational age, peak inspiratory pressure at 12 hours, and mean airway pressure at 10 days for 10-day-old neonates. These two regression analyses were applied prospectively to three new data sets totaling 238 neonates to test their predictive ability. Neonates were classified into low-, moderate-, or high-risk groups on the basis of their predicted probability of requiring oxygen supplementation at 28 days; low = probability of <25%, moderate = probability of 25% to 75%, and high = probability >75%. Although these populations were demographically distinct from the original group, the regression analyses performed well. The regression analysis for 12 hours of age classified 125 neonates at low risk of whom 9% required supplemental oxygen at 28 days, and the regression analysis for 10 days classified 141 neonates at low risk of whom 7% required supplemental oxygen. The 12-hour regression analysis classified 80 neonates at moderate risk of whom 33% required supplemental oxygen at 28 days and the 10-day regression analysis classified 49 neonates at moderate risk of whom 24% required supplemental oxygen. The 12-hour regression analysis classified 33 neonates at high risk of whom 64% required supplemental oxygen, and the 10-day regression analysis classified 48 neonates at high risk of whom 75% required supplemental oxygen. A graphic presentation of these two regression analyses permits easy bedside prediction of a neonate's risk of requiring supplemental oxygen at 28 days at the time of enrollment for clinical preventive or therapeutic trials for bronchopulmonary dysplasia and facilitates the calculation of required sample sizes.