Search in the Abstract Database

Abstracts Search 2019

P087 Urinary formate and glycine are associated with treatment response in patients treated with antibiotics for pouchitis

J. Segal*1, M. Sarafian2, J. I. Serrano Contreras2, A. Pechlivanis2, L. Braz1,2, Y. Siaw3, S. Clark1,2, E. Holmes2, A. Hart1,2

1St Marks Hospital, Gastroenterology, Harrow, UK, 2Imperial College London, London, UK, 3Hillingdon Hospital, Gastroenterology, Hillingdon, UK


Restorative proctocolectomy (RPC) is considered the preferred surgical choice for patient with ulcerative colitis (UC) who have failed medical therapy and in some patients with familial adenomatous polyposis (FAP). It has been shown through metabolic profiling of urine that CD patients have higher levels of formate and lower levels of hippurate and 4-cresol sulphate when compared with healthy controls. To date extensive metabolic profiling in RPC has yet to be studied. This study aimed to determine compounds found in urine that are associated with treatment response in patients that have been treated for pouchitis.


Patients with pouchitis were recruited from a single centre. Pouchitis was defined using the pouch disease activity index (PDAI) and pouchitis was considered when the score was ≥ 7. Response to antibiotics was defined as either a two points reduction in PDAI. Mid-stream morning urine samples were collected. Samples we stored at −80°C until analysis. 1H-NMR profile were recorded using the Bruker® Avance III 600MHz spectrometer, with a Samplejet 96 well autosampler. Standard 1-dimension NMR experiments with water suppression was performed at 300 K. All NMR spectra were automatically referenced to TSP at 0 ppm, phased and baseline-corrected on Topspin 3.2. Spectra were exported to Matlab for pre-processing. The full resolution 1H NMR spectra were imported into the SIMCA-P software package and multivariate data analyses were carried out. Once the NMR spectral regions related to the discrimination between two sample classes have been identified using supervised multivariate discriminant analysis, statistical total correlation spectroscopy (STOCSY) was applied. Metabolite assignment was performed by comparing chemical shifts, Jres coupling, and peaks multiplicity with information in databases (such as Human Metabolome DataBase, HMDB).


There were 21 patients. The median age of the cohort was 50 years (range 28–79). A total of 11 patients were on antibiotics and 10 patients were off antibiotics. Nine were responders. On multi-variate modelling there were significant differences found between responders and non-responders (CV-ANOVA p = 0.05). Significant spectral differences that corresponded to the multi-variate model correlated with Formate (8.84 PPM) Trigonelline (4.45PPM) and Glycine 3.57(PPM) all of which were higher in responders.


Trigonelline, formate and glycine may help differentiate patients with pouchitis who will respond to treatments vs. those that do not. It is currently unclear as to the mechanism as to why these metabolites are reduced in non-responders and further work is required to understand this and validate our findings.