Gut-brain Axis Dysfunction Underlies FODMAP-induced Symptom Generation in Irritable Bowel Syndrome

Jie Wu; Imke Masuy; Jessica R Biesiekierski; Heather E Fitzke; Chinar Parikh; Laurel Schofield; Hafsa Shaikh; Anisha Bhagwanani; Qasim Aziz; Stuart A Taylor; Jan Tack; Lukas Van Oudenhove


Aliment Pharmacol Ther. 2022;55(6):670-682. 

In This Article


Due to the fermentability and osmotic activity of FODMAPs, their ingestion can lead to intestinal distention through increases in luminal gas and water content. In patients with IBS, this process can induce GI symptoms, including bloating, diarrhoea, pain, cramps and flatulence, but the underlying mechanisms remain unclear. In this study, we investigated gut and neural responses to fructans (a key FODMAP), vs glucose and saline via MR-based abdominal and brain scanning. First, we confirmed our previous finding[6] of early (ie well within the first hour post-infusion) symptom induction by intragastric fructans infusion in IBS. Furthermore, we demonstrated that fructans increased small bowel motility and ascending colon gas and volume to a similar extent in IBS patients and HC, despite only IBS patients reporting increased levels of cramps, pain, flatulence and nausea, partially confirming earlier results of Major et al.[5] For the first time, we demonstrate differential brain responses to fructans (compared to glucose and saline) were observed between IBS and HC in pain-responsive regions including cerebellum, supramarginal gyrus, anterior and midcingulate cortex, insula and thalamus. These brain responses covaried more extensively with GI symptom responses in IBS compared to gut responses (only for colonic gas with bloating and cramps and colonic volume with cramps, and for the fructans vs saline comparison).

As anticipated, IBS patients reported increases in GI symptom ratings after fructans infusion compared to both glucose and saline control conditions, contrary to HC. The stronger increase in cramps in IBS patients vs HC after fructans is in line with our previous study,[6] which remarkably found a rapid increase in symptom ratings, with significance reached at 30 minutes post-infusion. Therefore, 1 hour of fMRI scanning should allow us to reach peak symptom intensity levels during scanning. In the current study, a significant rise in cramps, pain, flatulence and nausea following fructans compared to glucose was found in IBS, confirming our earlier findings, but it occurred more gradually compared to our previous study and continued to increase towards the end of the measurement. Although both glucose and saline were used as controls, it should be acknowledged that glucose has a calorie effect and can release GI hormones compared to saline, which may account for some of the differences from fructans. The difference in peak symptom intensity between our two FODMAP studies may be due to the higher concentration, as well as lower and fixed volume of the infusion used in the current study (Table S1). These changes compared to our previous study[6] were implemented to correspond with the study by Major et al,[5] but fructan was chosen over inulin as it is more common in the Western diet.[22] This could explain the difference in timing of symptoms, given the average degree of polymerisation (DP) and osmotic activity of fructans (DP = 3–10; 72 mOsmol/kg) is between that of fructose (DP = 1; 462 mOsmol/kg) and inulin (where long chain inulin types have a DP ≥23; 36 mOsmol/kg). As lower DP chains are more water soluble, they show more water retention properties than long chains and therefore we can expect greater osmotic effects induced by fructans than inulin in the small bowel. The increased small bowel motility after fructans could also accelerate the transit and prompt the arrival of fructans in the colon.[23] Furthermore, the intragastric infusion of the test solution compared to the oral drinking administration used by Major et al may have also accelerated the arrival of the test solution to the proximal colon. This is supported by the observed increase in signal intensity in the small bowel and ascending colon from 1 hour post-infusion of fructans (Figure S8).

Abnormal GI motility is a hallmark feature of IBS. However, the majority of previous studies have relied upon indirect measures of bowel motility or transit time. We used recent advances in MRI, which allow a direct assessment of small bowel motility to show fructans induced a similar increase in small bowel motility in both IBS and HC. This is consistent with accelerated small bowel transit time following ingestion of a fructose-sorbitol mixture[24] and increased small bowel motility following mannitol solution in HC.[25] The abovementioned supplementary small bowel signal intensity analyses indicate that the increased small bowel motility after fructans may be caused by distention induced by increased small bowel water content (Supplementary Material S2.3.3, Figure S8). We observed similar fructan-induced increases in gas and volume between IBS and HC up to 2 hours post-infusion. However, IBS patients showed an initial increase in gas in the ascending colon, which remained stable compared to a steadier increase in HC over the 2-hour period. This is contrary to a previous MRI study[13] which showed a lesser increase in ascending colon volume in IBS diarrhoea-predominant patients compared to HC following a mixed nutrient challenge in the early post-prandial phase. Supplementary analyses suggest that the increase in colon volume following fructans in our study was driven by gas rather than osmotic effects based on the change in signal intensity (Supplementary Material S2.3.3, Figure S9).

In IBS patients, we demonstrated that fructan-induced bloating and cramps were associated with changes in colonic gas, but not small bowel motility. This was consistent with the findings by Major et al[5] that colonic gas, but not small bowel water was linked to GI symptom responses. A previous study also showed that symptoms following 20 g lactose ingestion in Chinese IBS patients were associated with both the presence of rectal hypersensitivity assessed using barostat and gas production measured by a hydrogen breath test.[26] Yang et al found that the likelihood of having GI symptoms was related to the dose of lactose ingestion and hydrogen gas production.[27] Taken together, despite differences in osmotic effects and other properties, all FODMAPs are (variably) poorly absorbed, thereby undergoing colonic fermentation, which in turn leads to gas production playing an important role in symptom generation.

Previous studies showed small differences in symptom generation after oral administration of various FODMAPs despite their differences in osmotic activity and other properties.[28,29] A common characteristic of the different FODMAPs is fermentability in the colon leading to gas production, with the latter having been shown to be associated with symptom responses in the abovementioned studies (using hydrogen breath tests). These results are in line with our MRI-based finding of an association between colonic gas production and symptom responses, but our brain results indicate the additional role of visceral hypersensitivity reflected in altered brain response patterns. We also found that colonic volume is associated with cramps, indicating that distension by colonic contents may also be relevant to symptoms.

In addition to increased motility and luminal distension effects, other mechanisms at the level of the gut which cannot be measured using MRI, including mast cells, barrier dysfunction and intestinal microbiota may also be involved in FODMAP-induced symptom generation.[30,31]

At the brain level, pain is not encoded one specific cortical area. Instead, it is a complex non-linear process involving systems processing homoeostatic-afferent information as well as cognitive and affective circuits, and "top-down" descending modulatory pathways that send descending projections from the brainstem areas to the dorsal horn of the spinal cord.[32,33] Studies using fMRI have reported associations between visceral hypersensitivity and altered brain activation and/or connectivity patterns in IBS.[34–36] The current study, focusing on visceral hypersensitivity following FODMAP intake, showed differential activation patterns in several pain-responsive brain regions between IBS and HC. Compared to glucose, fructans infusion was associated with altered responses in cerebellum, supramarginal gyrus, postcentral gyrus, anterior and midcingulate cortex, anterior and middle insula, rolandic operculum, putamen and thalamus, in IBS vs HC. Similar results were found for fructans vs saline, including cerebellum, supramarginal gyrus, superior temporal gyrus, superior and middle frontal gyrus, anterior and midcingulate cortex, anterior and middle insula and rolandic operculum.

These results are partly in line with a meta-analysis of previous studies in IBS showing differential activation in the insula, anterior cingulate cortex and the thalamus following rectal balloon distention in IBS patients compared to HC.[12] Furthermore, fMRI studies showed cerebellar responses to (visceral) nociceptive stimulation.[37] Notably, these previous studies compare a test stimulus (eg rectal balloon distention) to the absence of this stimulus (eg deflated balloon), whereas our study compares a test stimulus (fructans infusion) to an "active" control stimulus (eg glucose and saline infusion). In addition to being far more ecologically and pathophysiologically valid compared to rectal balloon distension, fructan administration represents an indirect rather than a direct visceral pain stimulus, as the distention resulting from fructans reaching the bowel is thought to cause symptoms, rather than the ingestion of fructans per se. Nevertheless, the differential activation patterns in brain areas involved in pain regulation confirm the interaction between colonic distention induced by FODMAP intake and centrally regulated visceral hypersensitivity in IBS, thereby reinforcing the concept of IBS being a DGBI. Our finding that fructan-induced responses in most of these brain areas covaried with levels of virtually all fructan-induced GI symptoms further corroborates this interpretation, and identifies central mechanisms as the main driver of fructan-induced symptom generation in IBS for the first time. However, it should also be noted that visceral sensitivity has peripheral and central components. A limitation of this study is that we did not measure peripheral sensitivity, at the clinical level (eg by rectal barostat) nor at the molecular level (eg by quantifying peripheral immune responses to food components).

Interestingly, fructans infusion did not only increase activity in pain-responsive brain regions compared to glucose and saline in IBS, but also reduced activation in several of these brain regions over the post-infusion timecourse. Remarkably, consistently opposite patterns from those seen in HC were observed: where fructans induced higher brain activation compared to glucose and saline in HC, reduced activation following fructans compared to glucose and saline was seen in IBS, and vice versa. The decreased activation of the right anterior cingulate cortex following fructans infusion in IBS can be associated with reduced pain coping.[38] Alternatively, the decreased activation of the anterior cingulate cortex in IBS patients can be interpreted as dysfunction of opioid-rich descending modulatory pain pathways, which has been linked to pain chronification.[39] Furthermore, imaging studies have shown that descending pain pathways can be altered by emotional state or negative emotion.[40] Hence, there might be a link between psychological state, reduced activity in the anterior cingulate cortex and increased symptom responses in IBS. However, further research to verify this interpretation is warranted. Interpretation of these differences in the direction of the brain responses between IBS patients and HC is complicated, since it remains unknown whether the brain activity observed in the current study is driven by inhibitory or excitatory neurons. Finally, it should also be noted here that, due to the very different nature of the stimuli used in the current vs previous studies (single nutrient infusion vs repeated phasic balloon distension), different fMRI design and analysis methods are required (pharmacological MRI vs classic block- or event-related designs), further increasing difficulties with direct comparisons of studies.

The anterior cingulate cortex and the anterior insula are often jointly activated.[41,42] Accordingly, a reduced activation of the left anterior and middle insula following fructans infusion compared to glucose and saline was found in IBS, whereas opposite activation patterns were seen in HC. The insula plays an important role in processing and integrating viscerosensory and, more broadly, interoceptive information.[42,43] Therefore, reduced activation in IBS might indicate a disturbed processing of interoceptive information. However, our results do not allow us to identify whether the underlying mechanisms lie in the brain, or in the gut, in the absence of measurements of peripheral mechanisms which may be driving aberrant interoceptive signalling, such as localised immune responses which have recently been shown to play a role in post-infectious IBS.[44] On the other hand, IBS patients displayed increased activation of the right insula, which is involved in the integration of the emotional and interoceptive state. Asymmetric activation of the insula has been reported previously, with increased activation of the left insula in positive emotional conditions and activation of the right insula in negative emotional conditions.[43,45] Again, these findings support the link between emotional state, symptom perception and brain responses.

Additionally, increased activation of the right supramarginal gyrus was observed in IBS patients. Considering the role of the supramarginal gyrus in pain regulation and specifically attention to pain,[46] we can speculate that IBS patients are more attentive to symptom responses, which may in turn lead to increased symptom reporting.

Some limitations of our study should be addressed. First, only female subjects were included in this study. Although IBS is a disorder with female predominance, our results cannot be generalised to both sexes. Second, based on the results of the current study and the previous study by Major et al.,[5] it is possible that the peak of symptom intensity lies outside the time frame of brain scanning and consequently, larger differences in activity of pain-responsive brain regions may have been missed in the current study protocol. Nevertheless, we did observe within- and between-group differences in symptom and brain responses after fructans well within 1 hour post-infusion. While a longer fMRI measurement would have been desirable, it was practically and methodologically not feasible. Longer scanning periods challenge the subjects' tolerance, inevitably leading to artefacts, and are seldom used in pharmacological MRI studies to avoid excessive influence of scanner drift. Importantly, to acquire the abdominal scans, brain scanning needed to end as pharmacological MRI analysis does not allow interruption of brain scanning. Hence, our paradigm represents the best compromise to capture both gut and brain responses (both pre- and post-infusion) within subjects' tolerance range. Third, we acknowledge there might be putative differences between IBS subtypes.[47] However, IBS diarrhoea-predominant patients do not show a more favourable response to a low FODMAP diet compared to IBS constipation-predominant patients in published clinical trials[48,49] and a meta-analysis,[50] so we opted for a mixed sample to increase generalisability of our findings towards the entire IBS patient population. Our study was not adequately powered to analyse IBS subtypes groups separately. Studies in large sample sizes are warranted to address this issue. Fourth, the sample size of our study was rather small, but adequately powered for the hypothesised group-by-condition interaction effect based on our previous data. Nevertheless, our findings should be confirmed in a larger sample. Finally, given that a big dose of fructans was chosen in our study, our findings cannot be generalised to smaller doses of dietary fructans.

To conclude, we demonstrated increased symptom responses to fructans infusion in IBS. Fructans increased small bowel motility and ascending colon gas and volume to the same extent in IBS and HC. Colonic gas responses to fructans vs saline (but not vs glucose) were associated with bloating and cramps in IBS. No associations were found with small bowel motility responses. Differential responses to fructans infusion (vs glucose and saline) were found in IBS patients vs HC in pain-responsive brain regions, including the cerebellum, supramarginal gyrus, anterior and midcingulate cortex, insula and thalamus. Associations between these brain responses and virtually all GI symptoms were found in IBS. These fMRI observations provide direct evidence that dysfunction of the gut-brain axis may underlie FODMAP-induced symptom generation in IBS, more specifically by showing a strong link between aberrant brain responses and fructan-induced symptom generation; they may represent the central readout of hypersensitivity to normal FODMAP-induced gut responses. Future research should further elaborate on these findings to unravel the exact implications of these findings for IBS pathophysiology and treatment, but targeting hypersensitivity by pharmacological, dietary and/or psychological therapies (eg exposure-based cognitive behavioural therapy) may constitute promising avenues.