Early-onset Sepsis and Early Antibiotic Exposure Among Neonates with Complex Congenital Anomalies

Background: Early-onset sepsis (EOS) is associated with significant morbidity and mortality among neonates, and as a result clinicians have traditionally had a low threshold to initiate antibiotics for suspected sepsis. However, in the current era, EOS rates are low, and neonatal antibiotic exposure is now known to be associated with adverse outcomes. No published guidelines for antibiotic initiation exist that are tailored to neonates with complex congenital anomalies. Further, rates of EOS are unknown in this population, and physiologic instability due to their underlying conditions may be indistinguishable from infection. We aimed to measure rates of EOS and early antibiotic exposure among infants with congenital anomalies born in the Special Delivery Unit (SDU) at the Children’s Hospital of Philadelphia.

Methods: Retrospective study of infants born in the SDU between 2009 and 2020. EOS was defined as a positive blood culture within 7 days of life associated with antibiotic treatment for >5 days. We analyzed EOS and early antibiotic exposure rates overall, by year, and by admission unit: neonatal intensive care unit (NICU) or cardiac intensive care unit (CICU). Analysis included summary statistics, chi-squared analysis, and linear regression.

Results: Between 2009 and 2020, 4,353 neonates were born in the SDU. The rate of EOS was 0.57% (n=25), and did not significantly differ by unit (p=0.04). The most common underlying congenital anomalies among those with EOS were congenital diaphragmatic hernia and hypoplastic left heart syndrome (Fig. 1a). E. coli was the most common pathogen implicated. Overall, neonates in the NICU were more likely to experience early antibiotic exposure (antibiotics initiated within 72 hours of birth) compared to those in the CICU (49% vs 24%, p < 0.01) (Fig. 1b, Fig. 2). Patients in the NICU with early antibiotic exposure were more likely to be premature (59% vs 25%, p < 0.01) and weight less than 1500g at birth (10% vs 2.5%, p < 0.01) than those in the CICU (Fig. 1b). Rates of antibiotic initiation over time decreased in the NICU (p=0.001), while the CICU demonstrated an increase (p=0.024) (Fig. 2).

Conclusion: Despite low rates of EOS in neonates with complex congenital anomalies, rates of early antibiotic exposure remain high and vary between different units and diagnostic groups. Clinical concern for sepsis is likely influenced by prematurity, severity of illness, diagnosis, and other risk factors. Local unit guidelines may have influenced trends in antibiotic use over time. Development of targeted algorithms and stewardship strategies tailored to underlying diagnoses could help to decrease unnecessary antibiotic exposure.