Exploring the Gut Microbiome and its Potential Association with Bacteremia in Pediatric Hematopoietic Cell Transplant Patients
Katharine Morrison, Jesse Blumenstock, Ahmed Moustafa, Ceylan Tanes, Kyle Bittinger, Brian T Fisher, Caitlin W ElgartenAbstract
Background
Patients receiving hematopoietic cell transplant (HCT) frequently develop blood stream infections (BSIs) during and after the neutropenic period following transplant. BSIs may lead to serious complications such as sepsis and mortality. Recent studies have demonstrated that increased microbiome dysbiosis and lower Shannon diversity were indicative of later BSIs in adult HCT recipients. The aim of this study is to understand the associations between the gut microbiome in pediatric HCT patients and development of BSIs. This insight could inform the development of prophylactic strategies targeting the pediatric gut microbiome, aiding in the prevention of BSI and improving outcomes for HCT recipients.
Methods
We conducted a prospective, longitudinal observational study of pediatric patients undergoing HCT for any indication at a single center. Stool samples were collected starting at admission through day +100 post-HCT. After DNA extraction from samples, shotgun libraries were generated and sequenced. Bacterial taxa were quantified by kraken2. Stool microbiome layout was summarized by alpha diversity (richness, Shannon) and a dysbiosis score defined based on the median distance to age-similar healthy controls. These scores were compared at baseline across those who did and did not develop a BSI by day +100.
Results
Sixty-two patients had a pre-HCT stool sample and were included in this analysis; demographic characteristics are shown in table 1. Of those, 17 experienced BSI, 8 of which were classified as mucosal barrier injury (MBI) BSI based on National Healthcare Safety Network definitions. Those who developed BSI demonstrated a lower mean richness (93.6 versus 113.1), lower Shannon index (2.90 versus 3.24) and higher dysbiosis score (0.81 versus 0.76) at baseline compared to who did not, but none of these comparisons reached statistical significance. Qualitatively, we also did not observe consistent differences/evolution in any microbiome index immediately prior to a BSI (demonstrative figure shown).
Conclusions
The baseline microbiome in pediatric HCT recipients was similar in those who develop BSI and those who do not, at least as measured by the three variables dysbiosis, richness and Shannon diversity. These data suggest that we need to look beyond traditional summary measures of microbiome disturbance to identify measures that are able to predict future bacteremia events such as relative or absolute abundance of specific taxa, genera, or even species.