174. Shotgun Metagenomics and Colonization by Antibiotic-resistant Bacteria in Pediatric Hematopoietic Stem Cell Transplant Recipients

BACKGROUND: Bacteremia in hematopoietic stem cell transplant (HSCT) recipients most frequently arises from gut bacterial translocation and is associated with a higher mortality if the organism is antibiotic-resistant. We sought to determine the impact of prior antibiotic exposure on antibiotic resistance genes (ARGs) in the gut metagenomes of HSCT recipients to inform future infection prevention strategies. METHODS: We performed shotgun metagenomic sequencing of fecal samples collected during the transplant hospitalization from children (< 18 years of age) undergoing HSCT at Duke University between 2015 and 2018. Host-decontaminated sequencing reads were aligned to the Comprehensive Antibiotic Resistance Database. We used a negative binomial regression model to determine the impact of recent therapeutic antibiotic exposure on the number of ARGs prior to HSCT. RESULTS: Median age of the 77 children included in these analyses was 4.8 years, and 58% were male. Hematological malignancy was the transplant indication for 42% of children, and 87% of transplants were allogeneic. In the 654 longitudinal samples, we identified 926 unique ARGs, conferring resistance to 31 classes of antibiotics. The median number of ARGs per sample was 24 (interquartile range: 13, 49). The most common ARGs detected were dfrF (conferring resistance to trimethoprim), tetO (tetracyclines), and tetW (tetracyclines), each detected in >65% of samples. Of the 66 children with fecal samples collected prior to HSCT, 70% of children received therapeutic antibiotics in the 2 weeks prior to enrollment. Accounting for transplant indication, sex, and age, the incidence of ARGs was 47% higher in children who received recent therapeutic antibiotics (incidence rate ratio 1.47; 95% CI (1.03–2.13); p=0.04). CONCLUSION: ARGs are commonly found in the gut metagenomes of pediatric HSCT recipients prior to HSCT and are associated with recent receipt of therapeutic antibiotics. Future directions for this dataset include determining the ability of ARGs in the metagenome to predict clinical outcomes, including mortality and infections. Understanding the colonization and acquisition of ARGs could inform infection prevention strategies and empiric therapies and lead to improved infectious outcomes in these high-risk patients. DISCLOSURES: All Authors: No reported disclosures