Microbiota Predict Infections and Acute Graft-Versus-Host Disease after Pediatric Allogeneic Hematopoietic Stem Cell Transplantation.

BACKGROUND: Despite intensive prophylactic and pre-emptive measures, infections remain an important cause of morbidity and mortality in pediatric recipients of allogeneic hematopoietic cell transplantation (allo-HCT). Disruption of the gut microbiota has been linked to clinical outcomes after adult allo-HCT. This study evaluated whether these or differing microbiota disruptions or signatures were associated with risk of infection in pediatric allo-HCT.

METHODS: In a prospective observational study, fecal samples from 74 children were collected prior to conditioning and at the time of neutrophil recovery and profiled by means of 16S ribosomal rRNA sequencing. The associations between microbiome signatures and infections or acute graft-versus-host disease (aGVHD) were examined using Cox proportional-hazards analysis.

RESULTS: Previously associated indices of microbiome disruption in adults, including diversity and butyrate producer frequency, did not predict infection risk in pediatric allo-HCT. Unique microbiota signatures were associated with different infections or aGVHD. A ratio of strict and facultative anaerobes (e.g. Lachnoclostridium, Parabacteroides, Clostridium spp.) prior to conditioning predicted likelihood of bacteremia (cox hazards ratio 3.89) in first year post HCT. A distinct ratio of oral (e.g. Rothia, Veillonella spp.) to colonic anaerobes (e.g. Anaerobutyricum, Dorea, Romboutsia spp.) at neutrophil recovery predicted likelihood of bacterial infections (cox hazards ratio 1.81) and viral enterocolitis (cox hazards ratio 1.96) through first year post transplant.

CONCLUSIONS: Interactions between medical interventions, pediatric hosts and microbial communities may be responsible for these consistent microbiota signatures that predict infections. A future multi-center investigation will be needed to demonstrate whether these ratios can be generalized to other pediatric cohorts.