P2C246: Modeling the Impact of Reducing Air Pollution on Acute Bronchopulmonary Dysplasia Exacerbations Among Preterm Infants After NICU Discharge

Background: Among infants with bronchopulmonary dysplasia (BPD), emergency department (ED) visits and hospital readmission for BPD exacerbations are common after initial NICU discharge. We previously reported an association of higher particulate matter (PM2.5) exposure with medically attended acute respiratory illness specifically among infants with grade 1 (mild) BPD. The World Health Organization (WHO) guidelines state that no safe level of air pollution exists. The impact of reducing PM2.5 exposure on health care use in infants with BPD is unknown.

Methods: In a dataset of infants with BPD, born between 2010-2020, and discharged home from a single hospital system to a residential address in the Philadelphia region, we analyzed associations of mean PM2.5 in the year after discharge with medically attended acute respiratory illness identified using ICD 9/10 codes. We geocoded each address and used a machine-learning model to estimate daily residential block group PM2.5 exposure. Models quantified average marginal effects (AME) of decreasing annual mean PM2.5 exposure by 1 μg/m3 with medically attended acute respiratory illness, adjusting for age at NICU discharge, year, sex, race/ethnicity, insurance, BPD severity (2019 NRN criteria), and census tract neighborhood deprivation. Secondary analyses evaluated ED visits and admissions separately.

Results: We evaluated 378 infants, 52% Black and 62% publicly insured (Table 1). Substantial regional variation in PM2.5 exposure existed; mean (SD) annual PM2.5 exposure was 8.70 (1.04) μg/m3. Among all infants, PM2.5 exposure did not differ significantly between infants with and without medically attended acute respiratory illness. Among infants with mild BPD, PM2.5 was significantly higher in infants with medically attended acute respiratory illness (Table 1). While overall, we did not detect significant associations of reducing PM2.5 exposure by 1 μg/m3 with overall medically attended acute respiratory illness, ED visits, or inpatient readmissions, among infants with mild BPD, reductions were associated with decreases in medically attended acute respiratory illness (AME -9.64%, p = 0.017) and decreased ED visits (-5.56%, p = 0.027), but not inpatient readmission (AME -3.71%, p = 0.347).

Conclusion: Per 1 μg/m3 decrease in PM2.5 annual mean concentration, our model revealed that infants with mild BPD could have a nearly 10% reduction in medically attended acute respiratory illness in the year after NICU discharge. Policies aimed at lowering annual PM2.5 exposure have important implications for healthcare utilization among infants with mild BPD.