Microbiome-driven IBS metabotypes influence response to the low FODMAP diet: insights from the faecal volatome

Abstract

Background: Irritable bowel syndrome (IBS) is a common and debilitating disorder manifesting with abdominal pain and bowel dysfunction. A mainstay of treatment is dietary modification, including restriction of FODMAPs (fermentable oligosaccharides, disaccharides, monosaccharides and polyols). A greater response to a low FODMAP diet has been reported in those with a distinct IBS microbiome termed IBS-P. We investigated whether this is linked to specific changes in the metabolome in IBS-P.

Methods: Solid phase microextraction gas chromatography-mass spectrometry was used to examine the faecal headspace of 56 IBS cases (each paired with a non-IBS household control) at baseline, and after four-weeks of a low FODMAP diet (39 pairs). 50% cases had the IBS-P microbial subtype, while the others had a microbiome that more resembled healthy controls (termed IBS-H). Clinical response to restriction of FODMAPs was measured with the IBS-symptom severity scale, from which a pain sub score was calculated.

Findings: Two distinct metabotypes were identified and mapped onto the microbial subtypes. IBS-P was characterised by a fermentative metabolic profile rich in short chain fatty acids (SCFAs). After FODMAP restriction significant reductions in SCFAs were observed in IBS-P. SCFA levels did not change significantly in the IBS-H group. The magnitude of pain and overall symptom improvement were significantly greater in IBS-P compared to IBS-H (p = 0.016 and p = 0.026, respectively). Using just five metabolites, a biomarker model could predict microbial subtype with accuracy (AUROC 0.797, sensitivity 78.6% (95% CI: 0.78-0.94), specificity 71.4% (95% CI: 0.55-0.88).

Interpretation: A metabotype high in SCFAs can be manipulated by restricting fermentable carbohydrate, and is associated with an enhanced clinical response to this dietary restriction. This implies that SCFAs harbour pro-nociceptive potential when produced in a specific IBS niche. By ascertaining metabotype, microbial subtype can be predicted with accuracy. This could allow targeted FODMAP restriction in those seemingly primed to respond best.

Funding: This research was co-funded by Addenbrooke's Charitable Trust, Cambridge University Hospitals and the Wellcome Sanger Institute, and supported by the NIHR Cambridge Biomedical Research Centre (BRC-1215-20014).

Keywords: Irritable bowel syndrome; Low FODMAP diet; Metabolome; Metabotype; Microbiome; Short chain fatty acids; Volatile organic compounds.

Fig. 1

Flow chart outlining study design. At each time point, the samples available for analysis are reported for both the index study conducted by Vervier et al. (2022) and the current metabolomic study. The slight increase in drop out in the metabolomic study reflects the inability to perform metabolomic analysis due to biological sample depletion (FODMAP, fermentable oligo-, di-, mono-saccharides and polyols; IBS-SSS, IBS symptoms severity scale).

Fig. 2

a: Self organising heatmap clustered according to microbial subtype showing the distribution of metabolites during the consumption of the baseline/habitual diet. b: Correlation plot demonstrating relationships between individual volatile organic compounds (VOCs). Positive inter-VOC correlations are observed between the VOCs characterising each metabotype, whereas negative inter-VOC correlations are observed between metabotypes. (N = 56; 28 IBS-P, 28 IBS-H).

Fig. 3

Two-dimensional loading plots following unsupervised principal component analyses demonstrating the degree of similarity/differences between groups. a) IBS-P and their controls; b) IBS-H and their controls; c) IBS-P and IBS-H and all controls (note how the degree of overlap is higher between IBS-H and controls vs IBS-P and controls).

Fig. 4

Left: Heatmap demonstrating the key differences between the IBS-H and IBS-P subtypes at baseline. Right: Receiver operator characteristic demonstrating a microbial class (IBS-H/IBS-P) prediction model based on 5 of the most abundant and discriminant volatile organic compounds.

Fig. 5

IBS-P group moving from baseline to completion of a 4-week low-FODMAP diet (LFD). N = 18, paired data. a: Volcano plot illustrating the dynamic changes in VOC abundance. Significance parameters set to require—log-2-fold change AND significant unadjusted p value (alpha <0.05) on Wilcoxon signed rank test. b: Box plots illustrating the significance of the dynamic changes in SCFA and branched SCFA abundance (Wilcoxon signed rank test).

Fig. 6

Differences in SCFA abundance at baseline and after completion of a 4-week low-FODMAP diet. Unadjusted p values and false discovery rate adjusted q values reported for non-parametric ANOVA. Subsequent significant differences between groups identified on post-hoc analysis are highlighted using the asterisk. N = 78 (39 IBS [18 IBS-P, 21 IBS-H], 39 controls [19 Control-P, 21 Control-H]).