Should Vitamin A Injections to Prevent Bronchopulmonary Dysplasia or Death Be Reserved for High-Risk Infants?: Reanalysis of the NICHD Neonatal Research Network Randomized Trial.

Objective To determine whether infants at higher risk of bronchopulmonary dysplasia (BPD) or death benefit more from vitamin A therapy than those at lower risk. Study design Post-hoc reanalysis of a landmark phase-3 randomized controlled trial conducted from January 1996-July 1997 at 14 university-affiliated neonatal intensive care units in the United States. Data analysis performed October 2019-October 2020. Infants born weighing 401-1000 g and receiving respiratory support at 24 hours of age were assigned to intramuscular vitamin A 5000 IU or sham procedure 3 times weekly for 4 weeks. The primary outcome was BPD, defined as use of supplemental oxygen, or death at 36 weeks’ postmenstrual age. An externally validated model for predicting BPD or death was used to estimate the risk of these outcomes for each infant. Results As previously reported, 222/405 (54.8%) infants assigned vitamin A therapy and 248/402 (61.7%) in the control group developed BPD or died (RR=0.89 [95% CI: 0.80-0.99]; RD=-6.9% [-13.0 - -0.7%]). Predicted individual risks of BPD or death ranged from 7.1 to 98.6% (median 61.5%; mean 60.9%). The effect of vitamin A therapy on BPD or death depended on infants’ risk of the primary outcome (p=0.03 for interaction): for example, RR=0.73 (RD=-14.5%) for infants with 25% predicted risk and RR=0.96 (RD=-1.0%) for infants with 75% risk. There was no difference in reduction of vitamin A deficiency across risk groups. Conclusion Contrary to expectation, the effect of vitamin A therapy on bronchopulmonary dysplasia or death was greater for lower-risk than higher-risk infants.