THE PREOPERATIVE FAECAL LIPODOME BUT NOT MICROBIAL DIVERSITY PREDICTS POST-OPERATIVE ILEUS IN ELECTIVE COLORECTAL SURGERY
James Lewis1, Sam Mason*1, Petra Paizs1, David J. Brinkman2, Simon Cameron1, Alvaro Perdones-Montero1, Ara Darzi1, Zoltan Takats1, Wouter de Jonge2, James M. Kinross1
1Surgery & Cancer, Imperial College London, London, United Kingdom; 2Gastroenterology, Academic Medical Centre, Amsterdam, Netherlands
Background:
There is increasing evidence that microbial-host interactions modulate postoperative bowel function and therefore may have a role in postoperative ileus (POI). The analysis of the faecal metabolome provides a unique insight into host-microbial interactions which may allow mechanistic investigation and clinical risk stratification. The aim of this pilot study was to identify if the pre-operative stool lipidome is predictive of POI and to explore possible causes of its variation.
Methods:
This was a cross-sectional study of the preoperative faecal lipidome and microbiome of patients recruited to the multicentred RCT SANICS II (Peters et al, 2018) in Denmark and the Netherlands. Adult patients were included if they underwent elective colorectal resection and were excluded if they had previous surgery likely to disrupt the microbiome or had medication to modulate acetylcholine metabolism. Fresh pre-operative faecal samples were biobanked and metabolites with a mass:charge ratio of 120-600Da were analysed by Laser-assisted Rapid Evaporative Ionization Mass Spectrometry (LA-REIMS) in positive and negative mode using a Xevo G2-XS QToF mass spectrometer (Waters Corporation). LA-REIMS represents a high throughput analytical platform with no requirement for sample preparation and it can be deployed at the point of care. Data were analysed using multivariate statistical methods including Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA) in SIMCA v14.1 (Umetrics). 16S rRNA sequencing data were analysed using MicrobiomeAnalyst (www.microbiomeanalyst.ca).
Results:
42 patients (78% male, mean age 68) were included in the analysis with 90.5% of resections performed for malignancy. 114 spectra were generated across positive and negative modes, with 102 passing quality control. PCA demonstrates the pre-operative faecal lipidome was not significantly influenced by age, gender, BMI, malignancy status, smoking status, ASA grade, diabetes mellitus or tumour stage. In positive mode, an OPLS-DA model predicted the 23 patients who developed post-operative ileus following leave-one-spectrum-out cross validation (R2 0.24, Q2 0.132, AUC 0.9). There was no significant difference between the ileus and non-ileus patients in the Shannon diversity index of the microbiome (p = 0.09). LA-REIMS in negative mode had no predictive power for POI.
Conclusions:
The preoperative faecal lipidome can be detected using LA-REIMS and may predict the development of POI. The faecal microbiome does not appear to contribute significantly to this model, implying host-microbial co-metabolism rather than changes in faecal diversity contribute to the mechanisms of POI.
Back to 2019 Posters
