Methods
A retrospective chart review was performed on subjects (n=417) recruited from outpatient BPD clinics at the Children’s Hospital of Philadelphia and Johns Hopkins Children’s Center between 2008-2021. All subjects were born at less than 37 weeks-gestation and were diagnosed with BPD by NIH consensus definition,11 and were on home supplemental oxygen via nasal cannula; patients with tracheostomies or home ventilators were excluded. This study was approved by Johns Hopkins Institutional Review Board (Protocol #NA_00051884) and the Children’s Hospital of Pediatrics Institutional Review Board (IRB# 20-017614) and all caregivers were consented.
Children using supplemental O2 at initial hospital discharge were divided into four groups. Those who required ≤1/8 lpm of O2 (group 1), those who required >1/8 lpm of oxygen but <1/2 lpm of O2 (group 2), those who required ≥1/2 lpm O2 but <1 lpm O2 of oxygen (group 3), and those who required ≥1 lpm O2 of oxygen (group 4). Demographic data and clinical characteristics were collected by chart review. Birthweight percentile was corrected by gestational age.12 Race/ethnicity was obtained through caregiver report. Median household income was estimated using residential zip codes based on U.S. Census data.
Acute care usage and chronic respiratory symptom data was obtained via caregiver-completed questionnaires at outpatient visits when subjects were between 0 and 3 years of age.
Statistics: Chi square and ANOVA tests were performed to compare demographic data and clinical characteristics between oxygen groups (Table 1). Cox regressions adjusted for inter-site variation were performed to identify factors associated with earlier or later liberation from supplemental oxygen (Table 2). Associations between amount of oxygen at initial hospital discharge and acute care use/chronic respiratory symptoms were assessed for via logistic regression models adjusted for inter-site variation, potential demographic/clinical confounders, duration of oxygen use, number of questionnaires, and age of questionnaire assessments (Table 3). Results that were statistically significant had a P value less than or equal to 0.05. Stata IC 16 was used for data analysis.