Covariation of Neonatal Intensive Care Unit-Level Patent Ductus Arteriosus Management and In-Neonatal Intensive Care Unit Outcomes Following Preterm Birth

Abstract

To test the hypothesis that neonatal intensive care unit (NICU)-specific changes in patent ductus arteriosus (PDA) management are associated with changes in local outcomes in preterm infants. This retrospective repeated-measures study of aggregated data included infants born 400-1499 g admitted within 2 days of delivery to NICUs participating in the California Perinatal Quality Care Collaborative. The period 2008-2015 was divided into four 2-year epochs. For each epoch and NICU, we calculated proportions of infants receiving cyclooxygenase inhibitor (COXI) or PDA ligation and determined NICU-specific changes in these therapies between consecutive epochs. Generalized estimating equations were used to examine adjusted relationships between NICU-specific changes in PDA management and contemporaneous changes in local outcomes. We included 642 observations of interepoch change at 119 hospitals summarizing 32 094 infants. NICU-specific changes in COXI use and ligation showed significant dose-response associations with contemporaneous changes in adjusted local outcomes. Each percentage point decrease in NICU-specific proportion treated with either COXI or ligation was associated with a 0.21 percentage point contemporaneous increase in adjusted local in-hospital mortality (95% CI 0.06, 0.33; P = .005) among infants born 400-749 g. In contrast, decreasing NICU-specific ligation rate among infants 1000-1499 g was associated with decreasing adjusted local bronchopulmonary dysplasia (P = .009) and death or bronchopulmonary dysplasia (P = .01). NICU-specific outcomes of preterm birth co-vary with local PDA management. Treatment for PDA closure may benefit some infants born 400-749 g. Decreasing NICU-specific rates of COXI use or ligation were not associated with increases in local adjusted rates of examined adverse outcomes in larger preterm infants.