225 - Effectiveness of Vitamin A in Preventing Bronchopulmonary Dysplasia in Extremely Low Birth Weight Infants

Poster Board #: 225
Publication #: 2910-HP-QA.225

Background: Bronchopulmonary dysplasia (BPD) is one of the most common morbidities affecting as many as 40% of extremely low birth weight (ELBW) infants. Vitamin A had shown BPD reduction in ELBW infants from a randomized trial 20 years ago. However, the true effectiveness of vitamin A in the contemporary setting has been questioned as the rates of BPD did not change during 2011-2014, when there was a national shortage of vitamin A.


Objective: To determine the contemporary effectiveness of vitamin A for the prevention of BPD in ELBW infants among children’s hospitals using a propensity score matching approach.


Design/Methods: A national cohort of ELBW infants was identified from the Pediatric Health Information System from 2007-2010. Additional eligibility criteria were described under Table 1. Vitamin A treated infants were defined as those who received >= 3 doses, and high-users hospitals were defined as those who administered vitamin A to >=50% of their ELBW infants. Propensity score matching aimed to minimize bias by matching treated to untreated infants with similar propensity for receiving vitamin A based on patient characteristics (Table 2). Conditional logistic regression determined the association of vitamin A on BPD prevention.


Results: 2482 ELBW infants from 20 children’s hospitals were eligible in this study. Infants receiving vitamin A were more likely to be non-Hispanic Black infants (Table 1). Table 2 showed substantial improvement in the similarity of the treated and untreated groups after matching. There were no significant effects for mortality (Table 3). Vitamin A appeared to have significant benefit for BPD in the match that did not take hospitals into account (Table 3, Match 1), however, when matching was restricted to the same hospital, the observed benefit disappeared (Match 2). High-users hospitals had a trend towards lower BPD rates as shown in the match of untreated infants in high-users hospitals to untreated infants in low-users hospitals (Match 3).

Conclusion(s): Use of vitamin A was not associated with reductions in BPD after adjusting for treatment hospital. High-users hospitals overall had a trend toward lower BPD rates, but within these hospitals vitamin A continued to not be associated with lower BPD rates. These results suggest that there may be fundamental differences in care bundles that affect BPD at different hospitals that includes the use of vitamin A. Additional analysis using data from perinatal centers is underway to confirm these findings.

---

Authors/Institutions: Kuan-Chi Lai, The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States; Scott Lorch, The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States