Abstract
BACKGROUND: There is a limited literature describing the oral microbiome and its diagnostic potential in paediatric inflammatory bowel disease [IBD]. METHODS: We examined the dorsum tongue microbiome by V1-V2 sequencing in a cohort of 156 treatment-naïve children diagnosed with IBD compared to 102 healthy control children. Microbiome changes over time following treatment were examined in a subset of patients and associations between IBD diagnosis and dysbiosis were explored. RESULTS: Analysis of community structure of the microbiome in tongue samples revealed that IBD samples diverged significantly from healthy control samples [PERMANOVA p = 0.0009] and exhibited a reduced abundance of Clostridia in addition to several major oral genera [Veillonella, Prevotella and Fusobacterium species] with an increased abundance of streptococci. This dysbiosis was more marked in patients with severe disease. Higher levels of the potential pathobionts Klebsiella and Pseudomonas spp. were also associated with IBD. In terms of predicted functions, the IBD oral microbiome was potentially more acidogenic and exhibited reduced capacity for B vitamin biosynthesis. We used a machine learning approach to develop a predictive model of IBD which exhibited a mean-prediction AUC [area under the ROC curve] of 0.762. Finally, we examined a subset of 53 patients following 12 months of therapy and could show resolution of oral dysbiosis as demonstrated by a shift towards a healthy community structure and a significant reduction in oral dysbiosis. CONCLUSION: Oral dysbiosis found in children with IBD is related to disease severity and resolves over time following successful IBD treatment.