P077 Metabonomic profiling distinguishes Crohn’s perianal fistulas and idiopathic idiopathic (cryptoglandular) perianal fistulas: possible clues to underlying pathogenesis?

S. Adegbola1, M. Sarafian2, K. Sahnan1, P. Tozer1, N. Ding3, O. Faiz1, J. Warusavitarne1, R. Phillips1, E. Holmes2, A. Hart3

1St. Mark’s Hospital, Department of Colorectal Surgery, Harrow, UK, 2Imperial College London, Computational Systems Division, London, UK, 3St. Mark’s Hospital, Inflammatory Bowel Diseases Unit, Harrow, UK


The reasons why fistulas originate have been explained in idiopathic cases, with the cryptoglandular theory being widely accepted; however, in Crohn ́s disease, it is thought to involve interplay between microbiological, immunological and genetic factors. It remains unclear why fistulas persist. Novel investigative tools in systems biology are improving our understanding of pathogenesis, and one such tool is metabonomic profiling, whereby spectrometry is used to assess metabolic responses of complex systems in health and disease. These changes are mapped using analytical and statistical techniques. To the best our knowledge, these methodologies have not been applied to aid understanding of the pathogenesis of Crohn’s perianal fistula persistence.


Fistula tissue biopsy was obtained from the fistula tract of 30 patients with idiopathic perianal fistula and 20 patients with Crohn ́s anal fistula. Two analytical platforms were attempted to achieve broad metabolome coverage. Univariate (student t-test) and multivariate statistical data analyses were performed. From the latter principle component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA), models were built to find metabolites that could predict tissue samples from patients with either Crohn’s or idiopathic anal fistula. Metabolite putative identification was conducted by matching accurate mass:charge ratio (m/z) measurements of detected chromatographic features to theoretical value from in-house databases and on-line databases and previous publications.


Significant OPLS-DA predictive models (validated with CV-ANOVA p-value <0.05) differentiated metabolites from tissue samples from Crohn’s vs. idiopathic anal fistula patients using both metabonomic profiling platforms (HILIC/lipid profiling). A total of 22 differentiating metabolites were identified from the HILIC profiling and 19 from the lipid profiling analysis. Mapped pathway analysis was performed for metabolites identified using online database searches.


Forty-one differentiating metabolites were discovered belonging to various classes of lipids, amino acids and nucleotides. Pathways involved included amino acid metabolism, phospholipid metabolism and glycerolipid metabolism amongst others. Further work in larger numbers is required to validate these findings as well as cross correlation with microbiome work to understand the impact of host gut interactions in the pathophysiology of Crohn’s and idiopathic perianal fistulas.