Weight, habitual fibre intake, and microbiome composition predict tolerance to fructan supplementation

Abstract

Fructans are commonly used as dietary fibre supplements for their ability to promote the growth of beneficial gut microbes. However, fructan consumption has been associated with various dosage-dependent side effects. We characterised side effects in an exploratory analysis of a randomised trial in healthy adults (n = 40) who consumed 18 g/day inulin or placebo. We found that individuals weighing more or habitually consuming higher fibre exhibited the best tolerance. Furthermore, we identified associations between gut microbiome composition and host tolerance. Specifically, higher levels of Christensenellaceae R-7 group were associated with gastrointestinal discomfort, and a machine-learning-based approach successfully predicted high levels of flatulence, with [Ruminococcus] torques group and (Oscillospiraceae) UCG-002 sp. identified as key predictive taxa. These data reveal trends that can help guide personalised recommendations for initial inulin dosage. Our results support prior ecological findings indicating that fibre supplementation has the greatest impact on individuals whose baseline fibre intake is lowest.

Keywords: Microbiome; bioinformatics; dietary fibre; gastroenterology; inulin.

Figure 1.. Tolerance of inulin treatment in placebo-controlled trial.

A, Participant responses to questions about the Bristol scale score of their hardest, most frequent, and softest stool experienced during the treatment week. Mann-Whitney U test performed and no significant differences detected between treatment and placebo groups. B, Responses to questions asking whether the bars caused or reduced gastrointestinal (GI) discomfort, where 1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree. C, Responses to questions asking about specific GI side effects, where 1 = greatly reduced, 2 = reduced, 3 = did not affect, 4 = increased, and 5 = greatly increased. A-C, Mean and standard error shown. n = 40 participants (21 placebo, 19 prebiotic). Results of independent 2-group Mann-Whitney U tests shown, where * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 2.. Factors relating to diarrhea and stomach rumblings among study participants.

Scatterplots with Spearman correlation statistics for highest Bristol score (left) and stomach rumblings (right) with participant weight (A-B), habitual fiber intake as assessed by DHQ3 (C-D), and apparent change in fiber intake from Baseline week to Treatment week as assessed by ASA24 (E-F). n = 40 participants (C-D missing 1 prebiotic group participant, and E-F missing 3 placebo group participants due to surveys not being completed).

Figure 3.. Microbiome composition relates to GI discomfort and flatulence levels in response to prebiotic treatment.

A-C, Boxplots of the one ASV found to be different by any tolerance variables. SV 140, which mapped to Christensenellaceae R-7 group sp., was significantly different between low and high GI discomfort level groups during the baseline week (A) and treatment week (B), but not the follow-up week (C). Benjamini-Hochberg-corrected p-values from ALDEx2 GLM test controlling for study day shown. For GI discomfort, the low group refers to those who indicated a neutral response or disagreement to the statement that the treatment increased GI discomfort; the high group refers to those who agreed or strongly agreed. D, ROC curve of random forest models predicting flatulence response of prebiotic-treated subjects. The average (black) and iteration-specific (red) curves are both displayed. E, Average feature importance across iterations for the top 10 features. F, Changes in the centered log-ratio of bacterial taxa from week 1 to week 2 for the subjects by self-reported flatulence levels. Mean and standard error shown. AUC 95% CI 0.79–0.89, p(AUC > no information rate of 0.525) < 0.01. For flatulence, low group refers to those who reported “did not affect” or “increased”; high group refers to those who reported “greatly increased.” A-F, Prebiotic group only used in this analysis (n = 19).