P0777 - Variation of ductus arteriosus treatment in very low birth weight infants

Background
Patent ductus arteriosus (PDA) has long been recognized as a common neonatal condition, especially affecting very low birth weight (VLBW) infants. Although treatment is common, evidence to guide clinical management is limited. This study utilized a large US multicenter cohort of VLBW infants to describe PDA treatment variation by neonatal intensive care unit (NICU) classification.

Study Hypothesis
Clinical management of PDA in VLBW infants varies widely among US health care centers contributing to treatment heterogeneity.

Methods
Data were prospectively collected through Vermont Oxford Network (VON). The study cohort included 255,720 VLBW infants born 2012-2018 at 793 US health care centers; infants were inborn or admitted to the reporting facility within three days from birth and were never transferred. VLBW was defined as 401-1500g birth weight (BW) or 22-29 weeks gestational age (GA). Infants were excluded who died in the delivery room or prior to NICU admission, had congenital heart disease, received partial PDA treatment, or had incomplete hospital information. PDA treatment was categorized as pharmacological (ibuprofen, indomethacin, acetaminophen), invasive (surgical ligation, catheter closure) or combined (pharmacological and invasive). NICUs were classified into four types based on center-specific ventilation restrictions and/or invasive treatment capability reported by VON membership survey: (A) ventilation restrictions, and no invasive capability; (B) no ventilation restrictions, but no invasive capability; (C) surgery except cardiac requiring bypass; (D) cardiac catheterization and all surgery.

Results
Of the study cohort, 21% received PDA treatment: 83% pharmacological, 7% invasive, and 11% combined. Two percent of the study group were admitted at type A NICUs, 19% at type B, 50% at type C, and 29% at type D (Figure 1). Type A NICUs treated fewer infants (7%) compared to type B (13%), C (22%), and D (25%). The median treatment rate across all hospitals was 14% (IQR 6-23%), though treatment rates varied widely among and within NICU types (Figure 2). By definition, all type A and B NICUs only provided pharmacological treatment. Type C NICUs reported more pharmacological treatment (85%) compared to Type D (74%), but less invasive and combined (6 and 9% vs. 10 and 16%, respectively). Infants at type A NICUs, had higher BW (1,310g, IQR 1,145-1,430g) and GA (30 weeks GA, IQR 29-32 weeks) and lower prevalence of chronic lung disease (CLD) (10%). Infants at type D NICUs had lower BW (1,060g, IQR 785-1,310g) and GA (28 weeks GA, IQR 26-30 weeks) and higher prevalence of CLD (34%).

Conclusion
This study demonstrated very wide variation in PDA treatment of VLBW infants both across and within NICU classifications. Treatment variation was associated with infant characteristics as well as hospital treatment capability. PDA treatment heterogeneity and its potential effect on neonatal outcomes deserves further investigation.