LAUSR

BACKGROUND AND AIMS The cause of Crohn's Disease (CD) is unknown, but the current hypothesis is that microbial or environmental factors induce gut inflammation in genetically susceptible individuals, leading to chronic intestinal inflammation. Case-control studies of CD patients have catalogued alterations in the gut microbiome composition; however, these studies fail to distinguish if the altered gut microbiome composition is associated with initiation of CD or is the result of inflammation or drug treatment. METHODS In this prospective cohort study, 3483 healthy first-degree relatives (FDRs) of patients with CD were recruited to identify the gut microbiome composition that precedes the onset of CD and to what extent this composition predicts the risk of developing CD. We applied a machine learning approach to the analysis of the gut microbiome composition (based on 16S rDNA sequencing) to define a microbial signature that associates with future development of CD. The performance of the model was assessed in an independent validation cohort. RESULTS In the validation cohort, the microbiome risk score (MRS) model yielded a hazard ratio (HR) of 2.24, 95% CI [1.03, 4.84], and a p-value =0.04, using the median of the MRS from the discovery cohort as the threshold. The MRS demonstrated a temporal validity by capturing individuals that developed CD up to five years prior to disease onset (AUC > 0.65). The five most important taxa contributing to the MRS included Ruminococcus torques, Blautia, Colidextribacter, an uncultured genus-level group from Oscillospiraceace, and Roseburia. CONCLUSION This study is the first to demonstrate that gut microbiome composition is associated with future onset of CD and suggests that gut microbiome is a contributor in the pathogenesis of CD.