Mechanisms of hypersensitivity in IBS and functional disorders

Introduction

Visceral hypersensitivity is currently the leading hypothesis to explain irritable bowel syndrome (IBS) and other functional gastrointestinal disorders (FGIDs). After years of disappointing research trying to establish a correlation between the increased motor activity of the gut and the painful cramps felt by IBS patients, the theory that pain could be related to enhanced visceral sensitivity was raised. Ritchie,¹ in 1973, first reported that IBS patients were more sensitive than normal subjects to balloon distension of the colon. This observation of increased visceral sensitivity in IBS patients was confirmed by many researchers including Whitehead et al.², Mertz et al.³ and others. In agreement with the hypersensitivity of the colon found in IBS patients, intolerance to gastric distension was also documented in patients with functional dyspepsia (FD).^4–6

Visceral hypersensitivity: biological marker of FGID?

Mertz et al.³ proposed that ‘altered rectal perception is a biological marker of patients with IBS’ as it was identified in 94 of 100 IBS patients studied by a rectal distension by barostat. We collected rectal distension data by an electronic barostat in 164 patients (86 IBS, 26 painless constipation, 21 FD and 31 others with miscellaneous conditions),⁷ as also in 25 normal controls evaluated. As expected (Table 1), the pain thresholds to rectal distension were lower in patients with IBS when compared with control subjects; they were also lower than in patients with painless constipation, FD or other miscellaneous conditions. Assuming from the data in control subjects that 40 mmHg is the normal tolerance threshold, 90% of our patients with IBS were classified as suffering from visceral hypersensitivity. Severe hypersensitivity (less than 28 mmHg) was found in one third of our patients with IBS and was highly specific (95% specificity) for this condition. In our laboratory, the value of the barostat test for clinical diagnosis was reinforced by the negative predictive value of a pain threshold superior to 40 mmHg that reached 90%. These data on sensitivity, specificity and predictive values of the rectal distension test in patients with IBS are illustrated in Fig. 1.

In our experience, altered rectal perception in IBS is identified by the three following criteria: decreased pain threshold (found in about 90% of patients as indicated above); sensitization to repeated distensions (pain threshold at 40 and 41 vs 33 and 26 mmHg on the first and second sets of distensions in controls and in IBS patients, respectively); and pain referral to aberrant sites rather than the sacral area (present in 83% of IBS patients vs 10% of normal controls). These observations all support the concept of a dysregulation of the neural sensory system in IBS.

In other studies, we also tested the tolerance to gastric distension in 29 women and 10 men suffering from FD. Abnormal tolerance to fundic balloon distension was found in 77% of these patients.⁸

Visceral hypersensitivity: a peripheral mechanism

Visceral sensations are transmitted from the gut via afferent nerves travelling through the spinal cord to the brain, where their painful characteristics are then perceived. The brain response to balloon distension of the intestine can be measured by techniques such as positron emission tomography (PET) or by functional magnetic resonance imaging (fMRI). Despite their limitations, most of the studies with these instruments concur and show that the brain signal elicited in response to intestinal distension is greater in patients with IBS than in controls. This enhanced brain response could arise from an increased signal from the peripheral gut itself, the amplification of a normal intestinal signal during its travel through the spinal cord and brainstem or central signal amplification in the brain. Defects in the descending inhibitory mechanisms modulating pain transmission from the periphery to the brain could also be evoked.

In support of peripheral mechanisms involved in the enhanced pain transmission to the brain in IBS patients, the following arguments can be raised: (i) IBS and FD are known to occur after irritation of the gut by infectious agents^9–11; (ii) infiltration with inflammatory cells occurs in the gut mucosa or enteric plexuses of IBS patients^11–14; (iii) pain hypersensitivity is found exclusively in the visceral system and not in the somatic system in many IBS patients^2,15,16; (iv) rectal sensitivity to barostat distension can be increased by rectal instillation of glycerol¹⁷; (v) rectal sensitivity to barostat distension can be decreased by rectal administration of lidocaine¹⁸ and (vi) an abnormal intestinal reflex in the form of exaggerated motility can be obtained in IBS patients in response to repetitive distensions of the sigmoid colon at levels that are below the sensory threshold.¹⁹ All these data obtained in humans support the existence of an exaggerated signal from the periphery, that is the gut, although none of the arguments provides absolute proof of this concept. Whereas, it is possible to measure the brain response to intestinal distension (by PET or fMRI), there are no valid instruments to evaluate the signal at the level of the gut or of the spinal cord and brainstem in humans. Furthermore, it is unknown if neural dysregulation can occur at more than one site (e.g. in the gut and in the brain).

IBS and visceral perception: heterogeneity

Irritable bowel syndrome is currently recognized by symptoms and elimination of conditions that mimic it, such as lactose intolerance, microscopic colitis, coeliac sprue or Crohn's disease. Thus, IBS symptoms may be the phenotypic expression of other conditions. The pathophysiology of the symptoms in IBS may also be heterogeneous, some being hypersensitive, others having normal sensations. Visceral hypersensitivity itself may be heterogeneous, reflecting different pathophysiological abnormalities (e.g. peripheral dysregulation in some patients, but central dysregulation in others). These considerations could lead to the need for multiple, different therapeutic options.

Visceral hypersensitivity: frequent and symptomatic?

Patient heterogeneity is probably responsible for the discrepancy in the prevalence of visceral hypersensitivity, found in 90% of our IBS patients but in only 50% of the patients in other series. This does not invalidate the concept of visceral hypersensitivity in IBS. Hypersensitivity of the rectum is common in IBS, but not in other FGIDs such as functional constipation or FD (Table 1). Mertz et al.³ failed to identify specific symptoms that could relate to rectal hypersensitivity, apart from constipation. We also documented that visceral sensitivity was not correlated to the severity of the gastrointestinal (GI) symptoms, the quality-of-life status, the psychological index or types of psychological symptoms (depression, anxiety, etc.) identified by the SCL-90 questionnaire.²⁰

We also observed that gastric hypersensitivity to distension, present in 80% of our patients, could not be predicted by the age or sex of the patients, type of dyspepsia (ulcer type vs motor type) or by the therapeutic response to an anticholinergic drug.⁸

Rectal hypersensitivity appears to be a stable phenomenon as it remained unchanged and failed to correlate with the beneficial evolution of GI symptoms and quality-of-life obtained during treatment by group counselling psychotherapy^20,21 or SSRI antidepressant agents.^22–24 This finding is important as it confirms that visceral sensitivity is not just a reflection of the clinical symptoms and that improvement in GI symptoms can be achieved by factors not acting on visceral sensitivity. It is unclear as to when visceral hypersensitivity commences in the life of an individual, but it has been recognized in paediatric IBS patients.²⁵

While the severity of IBS symptoms was not correlated to rectal pain thresholds, the absence of correlation between the degree of the visceral hypersensitivity and the severity of the clinical symptoms does not negate the potential role of visceral hypersensitivity in IBS pathophysiology. Similarly, the severity of peptic ulcer disease or of oesophageal reflux is not well correlated with gastric acid secretion or oesophageal acid exposure.

Visceral sensitivity is therefore a heterogeneous condition in IBS patients: some are hypersensitive to rectal balloon distension, others are not. Current knowledge does not allow further differentiation of these groups of patients on the basis of symptom phenotype or the associated, probably heterogeneous, pathophysiology.

Visceral hypersensitivity: organ-specific or pan-intestinal?

Visceral hypersensitivity has been documented along the entire GI tract in FGID patients.^26–28 However, our observation⁷ that rectal sensitivity testing failed to identify patients suffering from FD led us to reconsider this paradigm. Therefore, we performed rectal and gastric distension tests in three groups of patients: patients with FD only (i.e. patients with discomfort or pain of the upper abdomen without any symptoms of the lower abdomen, suggestive of IBS); patients with IBS only (i.e. with symptoms restricted to the lower abdomen without any complaints of the upper abdomen) and patients with concomitant symptoms of the upper and lower abdomen.²⁹ As expected, 90% of patients were suffering inappropriate tolerance to distension at one of the two tests, but only approximately 33% of the patients were hypersensitive at both sites. As shown in Table 2, dyspepsia patients had gastric but not rectal hypersensitivity; IBS patients had rectal but not gastric hypersensitivity; while patients with dyspepsia + IBS symptoms were intolerant of distension at both sites. Heterogeneity of IBS patients is once again well shown here: some patients had hypersensitivity restricted to the colon, others to the whole GI tract. Fortunately, the topography of the visceral hypersensitivity can be predicted by the nature of the clinical symptoms (upper vs lower abdominal complaints).

Visceral hypersensitivity: stable or flexible?

During treatment with group counselling psychotherapy^20,21 or with SSRI antidepressants,^22–24 clinical improvement was obtained without changes in rectal pain thresholds. Does it mean that visceral hypersensitivity is a stable phenomenon that cannot be manipulated? When we treated patients with diarrhoea-predominant IBS (D-IBS) with amitriptyline,²⁰ clinical symptoms improved and rectal pain thresholds increased significantly (from 27.7 ± 1 to 33.7 ± 2 mmHg). The change in rectal sensitivity was well correlated (r = 0.71; P < 0.01) with the evolution of the GI symptoms. Fig. 2 shows, however, that patients appear to be divided into two groups: those who responded to amitriptyline (7/12 patients) and those who did not (5/12).

We also assessed the modulation of gastric hypersensitivity by the anticholinergic, hyoscine.⁸ Patients could be divided into two groups (Fig. 3): those whose intolerance to fundic distension could be improved with the anticholinergic (7/10 patients) and those whose could not (3/10 patients).

Overall, these data indicate that visceral sensitivity can be manipulated pharmacologically, but, once again, the response is heterogeneous. This heterogeneity in the response to a drug is recognized in large clinical trials of serotonergic agents. For example, about 50% of patients are responders to the 5-HT₄ agonist, tegaserod,^30–32 or to the 5-HT₃ antagonist, alosetron.^33–35

Is visceral hypersensitivity exclusive to the GI tract or due to a systemic condition?

Earlier studies^2,15,16 reported that IBS patients had tolerance to somatic pain that was normal or even increased. These data suggested that hypersensitivity was exclusive to the viscera. We found that tolerance to somatic pain, estimated by the cold-water tolerance test, was lower in FGID patients than controls³⁶ (Table 3). This somatic hypersensitivity in IBS patients has been confirmed by others.^37,38 Chang suggested that hypervigilance and concomitant fibromyalgia were predictors of somatic hypersensitivity in IBS patients.³⁸ In our laboratory, somatic hypersensitivity was a heterogeneous phenomenon identified in approximately 33% of our patients with FGID. No relationship was found between somatic pain perception and other factors, such as age, type of symptom (IBS or dyspepsia), severity of GI symptoms or SCL indices (total score or specific scores for somatization, depression, etc.).

Based upon these results on somatic and visceral sensitivity, we therefore propose that IBS patients constitute a heterogeneous population regarding their pain perception process: visceral hypersensitivity is common (80–90% of patients) and is accompanied in a significant proportion (30–50%) of these patients by somatic hypersensitivity, while some patients (10–20%) have normal pain sensation. Our work on visceral and somatic sensitivity generated a working hypothesis on the mechanisms for pain tolerance in IBS (Fig. 4): in patients with exclusive visceral hypersensitivity, enhanced brain signalling may originate from the gut itself or from amplification at the level of the spinal cord or the brain. However, in patients with concomitant somatic hypersensitivity, modulation of a peripheral signal at the level of the spinal cord or in the brain probably takes place to explain both hypersensitivities.

Conclusion

Visceral hypersensitivity is a frequent finding in IBS patients (50–90% patients in different series) and probably plays a role in the symptoms of these patients. However, IBS, as currently identified, remains still a heterogeneous condition. In IBS patients, visceral hypersensitivity can be present or absent, can be organ-specific or distributed to the entire GI tract, can be drug resistant or responsive, can be exclusive to the GI tract or accompanied by somatic hypersensitivity that may be generalized.

This heterogeneity is consistent with the existence of multiple pathophysiological disorders in IBS patients and therefore the potential for the need of multiple therapeutic options. Characterizing the individual patient for symptoms, physiological and genotypic characteristics may be needed to overcome the effects of the heterogeneity.

To minimize the contribution of this heterogeneity in FGID research, we must optimize the patient characterization (Fig. 5): the clinical phenotype (i.e. detailed and precise collection of the GI condition, as also extra GI symptoms, psychosocial conditions, therapeutic response to various interventions, etc.) must be characterized and correlated with the biological phenotype [i.e. pain perception processes (requiring new investigation tools to differentiate the contribution or of the gut or the medulla or the brain in this process), inflammation profile in the gut mucosa (including new markers recently described^11–14 and others to be discovered), motor activity, etc.] and the genotype (including currently suspected markers SERT³⁹ or GNβ3⁴⁰ etc. and others to be discovered).

Introduction: models for assessing central mechanisms in visceral sensation

Patients with IBS show enhanced perceptual responses (visceral hypersensitivity) to physiological visceral events, such as contractions and filling of the viscera⁴¹ as also enhanced sensitivity to certain types of experimental visceral stimuli, including distensions.^7,42,43 When comparing responses between humans and animal models, it is essential to highlight the difference between visceral sensitivity (referring to the sensitivity of the gut) and perceptual responses to visceral stimuli (referring to a subjective integrated response to a visceral stimulus). This poorly appreciated difference has confounded drug development strategies aimed at attenuating the enhanced perceptual responses in IBS patients. While the pharmaceutical industry has interpreted ‘visceral hypersensitivity’ as equivalent to increased sensitivity of primary visceral afferents and studies in animal models have yielded numerous molecular targets on these afferents, the human phenomenon of visceral hypersensitivity presents entirely different and complex challenges.

In humans, it is currently not possible to study true visceral afferent responses in a manner that would be comparable with single fibre responses to gut distension obtained in preclinical models. The great majority of human studies that evaluate the sensitivity of different parts of the GI tract to distension or chemical stimulation are using as a read-out the subjective ratings of stimulus intensity or unpleasantness. The only exceptions to this general approach are the more recent studies using different functional brain imaging techniques to assess the response of different brain circuits to gut stimuli. These central brain circuits are generally ignored by animal physiologists who study pseudo-affective reflexes, not subjective responses. In humans, visceral pain and discomfort are subjective, conscious experiences, which result from a central interpretation of the visceral afferent input that is influenced by context, memories, and emotional, motivational and cognitive factors. In principle, altered perception of visceral stimuli could result from activity changes in visceral afferent signal processing areas alone (reflecting increased visceral afferent input to the brain from the gut), from alterations in pain modulation circuits alone, or from variable combinations of these overlapping circuits. On the contrary, altered perceptual responses to gut stimuli can only be measured in human subjects. With the exception of higher non-human primates, it is presumed that animals do not have the brain circuitry for conscious perception of nociceptive stimuli. Thus, whereas preclinical studies use ‘pseudo-affective’ responses in intact animals as a read-out for responses to noxious (painful) stimuli or use in vitro preparations to eliminate the influence of the central nervous system (CNS) on afferent responses to stimuli altogether, perceptual responses to experimental gut stimuli in humans are highly processed by the CNS and may be poorly correlated with peripheral events in the viscera.

Evidence for a role of central mechanisms underlying visceral hypersensitivity in IBS

The brain has multiple ways to modulate the perception of afferent information, and this modulation is influenced by cognitive factors (e.g. attention and coping styles), the emotional state of the individual (e.g. fear, anxiety or anger) or memories of previous sensory events. Studies of both visceral and somatic pain have identified networks of brain regions involved in the cognitive, affective and motivational dimensions of the pain experience and associated autonomic nervous system responses (including subregions of the cingulate and prefrontal cortices, amygdala and hypothalamus).^44–46 These modulatory networks are generally observed in addition to those involved in primary pain processing (including the thalamus, insular cortex, and in some cases somatosensory S1 and S2 cortices). For example, studies examining patients with IBS when compared with healthy controls suggest that IBS patients may have increased activity during visceral inflations in sensory areas,⁴⁷ cortical areas associated with attentional processes and affective responses, and subcortical regions involved in affect, arousal and autonomic responses.^48–50

As discussed elsewhere in this issue, frequently overlapping mechanisms have been elucidated in animal models, which can result in enhanced responses to noxious stimuli in a variety of pain models. They include peripheral and central (spinal) sensitization,⁵¹ spinal bulbospinal pain amplification loops⁵² and descending inhibitory and facilitatory pathways.⁵³ In humans, it is likely that the conscious perception of peripheral changes at the gut level (including those resulting in peripheral and spinal sensitization) is always highly filtered and modulated by central mechanisms, including central pain amplification by cognitive and emotional factors.⁵⁴ In other words, while peripheral and central (spinal) sensitization may play a role in reported hypersensitivity in some IBS patients,^51,55,56 a significant component of the enhanced perception may also be due to central factors.

Central modulation: role of emotional context

It has been hypothesized that affective responses, including symptom-related anxiety, significantly alter perceptual responses to pain stimuli via multiple brain mechanisms including attentional, opioidergic and descending modulation networks.^46,57,58 Considerable progress has been made both on a preclinical and, more recently, on a clinical level to identify brain regions, circuits and mechanisms which play a role in the facilitation and inhibition of the subjective pain experience.^54,58.

Phillips et al. used a study paradigm involving viewing of emotional faces and non-painful oesophageal distension to evaluate the neural mechanisms underlying the effect of emotional context on visceral perception.⁵⁹ Brain responses to the oesophageal stimulation during either neutral or negative emotional context were evaluated. Activation within the right anterior insular cortex and bilateral dorsal anterior cingulated cortex (dACC) by the visceral stimulus was significantly greater while viewing fearful faces, compared with neutral faces.

In a second paradigm, the same group studied anxiety, discomfort and brain responses in another eight healthy male subjects during the same oesophageal stimulus while viewing faces with low, moderate or high intensity of fear expression. During the high-intensity fearful visual stimulus, significantly greater discomfort, anxiety and brain activation were observed, compared with the low-intensity fearful stimulus. Greater brain activation was seen predominantly in the left dACC and bilateral anterior insular cortex, the two regions most commonly reported in brain imaging studies of gut stimulation.⁶⁰ These findings clearly demonstrated the powerful effect of emotional context on the perceptual, emotional and brain response to an innocuous visceral stimulus.

Another study from Aziz's group assessed the modulatory role of attention on the brain responses to visceral (oesophageal) distension in seven healthy volunteers (six males).⁶¹ Brain responses to visual and oesophageal (non-painful balloon distension) stimuli were presented simultaneously, while subjects were asked to focus their attention on either the oesophageal or the visual stimulus, an established paradigm of selective attention. During another manipulation, subjects were asked to focus on a change in frequency of both stimuli, to simulate divided attention. Selective attention on the oesophageal stimulus was associated with activation of sensory (somatosensory cortex) and cognitive (dACC) networks, while selective attention on the visual stimulus activated the visual cortex. During the divided attention task, more brain regions in the sensory and cognitive domains were activated to process oesophageal stimuli, in comparison to those processing visual stimuli. These findings emphasize the importance of attentional processes in the modulation of sensory information from the body, and reflect the relatively greater biological importance placed on visceral sensation, compared with other sensory stimuli.

Evidence for relationship between hypervigilance, central arousal and visceral hypersensitivity

Hypervigilance can be produced experimentally in a condition of anticipated pain without delivery of the stimulus. Such hypervigilance may be associated with altered attentional and affective modulation of perception.⁶² The attentional modulation refers to an increased tendency to attend to the sensation and affective modulation refers to the increased aversiveness and threat from gut sensations.⁶² Experimental paradigms aimed at identifying the brain correlates of hypervigilance have shown this hypervigilance to be associated with activation of dorsomedial prefrontal cortex (dmPFC⁶³ amygdala) and dACC.⁶⁴ We recently reported greater activation of the mPFC, supragenual and infragenual ACC and amygdala in IBS patients when compared with healthy controls and patients with chronic gut inflammation, consistent with greater affective responses of the IBS group to the gut stimulus.⁴⁹ We and others have proposed a model in which increased activation of limbic and paralimbic circuits can lead to facilitation of noxious visceral input (via altered balance in the activity of inhibitory and facilitatory descending pain modulation systems), and hypervigilance to visceral sensations, via altered activity of central noradrenergic modulatory systems.^58,65,66 Recent evidence suggests that IBS patients may differ in the activation of brain regions involved in endogenous pain modulation.^49,67 Some of these responses, in particular the activation of endogenous pain facilitation systems, may also be observed in anticipation of distension.⁴⁸ These preliminary descriptive studies suggest that IBS patients show brain responses consistent with hyperresponsiveness to visceral stimuli in terms of vigilance, affect, arousal and perhaps sensory sensitization.

It is well known that repeated exposure to threatening or anxiety-provoking stimuli in the absence of aversive consequences leads to decreased attention and arousal, and consequently, reduced perception. Our group performed a study to examine longitudinal changes in visceral perception over the course of multiple rectal balloon distension sessions during a 12-month period and to identify the associated changes in brain activations.⁶⁸ The study was aimed at evaluating the role of hypervigilance and underlying brain circuit activation in the visceral hypersensitivity of IBS patients. While self-reported symptom severity and mood were relatively stable during the study period, the enhanced perception of experimental visceral stimuli observed at study entry gradually normalized over the repeated testing, and became similar to responses seen in healthy control subjects. Given the lack of change in subjective ratings of IBS symptom severity, it is likely that this reflects a process of habituation and decrease in hypervigilance to the repeated presentation of rectal stimuli. Comparison of brain responses over the study period confirmed this interpretation showing consistent activation of regions processing visceral afferent input over the 12-month period, but a reduced activation of and reduced functional connectivity within central networks concerned with vigilance and arousal during the second brain imaging session.

Conclusions

With the increasing use of functional brain imaging techniques to study brain responses to experimental visceral stimuli, such as gut distension, it has become possible to dissect out the different contributions made by ascending afferent input from the gut. Similarly, it is possible to study cognitive and emotional modulatory influences on the overall subjective experience of sensations from the viscera. Although few studies have assessed these factors in the context of gut stimulation, evidence from the somatic pain field strongly supports the powerful contribution of central mechanisms in the form of cortico-limbic-pontine brain circuits in the enhanced perception of peripheral stimuli. Regardless of additional abnormalities at the gut level resulting in peripheral and spinal sensitization, central sensory modulation is likely to be a crucial factor in the pathophysiology of ‘visceral hypersensitivity’ in IBS patients.

Introduction: inflammation and ‘functional disease’

Traditionally, functional GI disorders were defined by the absence of structural changes, but as investigative techniques become more sophisticated, it is increasingly recognized that there may be subtle microscopic abnormalities. As our understanding increases, the term ‘functional’ will become increasingly obsolete and disease will be described by mechanisms and not symptoms. Thus, cough (irritable lung) has been replaced by terms such as asthma, bronchitis and alveolitis, which describe pathology not symptoms. The gut is subject to numerous insults, which may result in inflammation, following which the tissues repair themselves so that it is often hard to see evidence of previous damage. While the reparative powers of the mucosa are exceptional, other structures are less likely to be restored completely to their original structure and function (Fig. 6). Thus, certain inflammatory cells such as Paneth cells and enterochromaffin cells (ECs) may remain abnormal for many years. The same is also likely to be true for intestinal muscle, interstitial cells of Cajal, and enteric nerves. Mucosal changes may not actually be directly responsible for changes in sensitivity, but they may be useful markers of abnormalities of deeper structures.

Time course of injury and repair

When assessing the effects of inflammation on gut sensitivity, it is important to recognize that there is a complex time course. The acute injury, which may be associated both with decreased and enhanced responses, is then followed by recovery and remodelling of injured tissue with resulting changes in sensitivity. Sensitization may be due not only to peripheral alterations in innervation and compliance, but also to a central sensitization or even desensitization, which arises as a result of chronic inflammation.

Regional differences in inflammatory injury

The nature of the insults and resulting injuries vary by location (Fig. 7). The oesophagus is protected by a substantial squamous epithelium, which provides a robust barrier to ingested toxins. The main injurious agent is refluxed gastric acid, bile and duodenal enzymes, which produce oesophagitis in which the depth of injury is limited to superficial layers, rarely extending to the muscle layer. The lining of the stomach is much thinner but well protected with mucus. The main toxic agent is again acid but also H. pylori, an unusual bacteria which has adapted to survive in the acidic gastric environment. Helicobacter pylori is a non-invasive organism which induces only a superficial gastritis confined to the mucosa. By contrast, the small bowel, whose main function is to absorb, is relatively poorly defended, with only a very thin lining and an extensive surface area. The main pathology is bacterial enteritis, which tends to induce mucosal injury. By contrast, in inflammatory bowel disease such as Crohn's disease, the injury may be very deep and transmucosal. Within the colon, the main pathologies are inflammatory bowel disease and diverticulitis. While ulcerative colitis (UC) is a superficial mucosal disease, diverticular disease induces deep inflammation and microabscess formation. The depth of inflammation and tissue injury in these different conditions is likely to impact on the subsequent alterations in innervation and hence visceral sensation.

While there are numerous animal studies demonstrating that both gastric and intestinal inflammation produce both acute and long-term hypersensitivity, evidence in humans is quite limited. This review will consider data on the effect of oesophagitis, H. pylori gastritis, inflammatory bowel disease including Crohn's and UC, diverticular disease and, finally, postinfective IBS.

Oesophagitis and hypersensitivity

It is widely recognized that patients with oesophagitis have increased sensitivity to heat, alcohol, acid and peppers, all of which activate the vanilloid receptor TRPV1 to generate a burning sensation, known as ‘heartburn’. While inflammation in oesophagitis can lead to deep ulceration and fibrosis, most patients with oesophagitis have inflammation limited to the subepithelial layer. Immunohistochemical staining in oesophagitis demonstrates alterations in innervation with a striking increase in the pan-neuronal marker PGP 9.5 and VIP.⁶⁹ This is mainly seen in inflamed mucosa, but lesser increases are seen in the expression of substance P (SP), changes which appear to resolve with healing after acid inhibition. Increased sensitivity to infused acid has long been used as a diagnostic test (Bernstein test) when attempting to identify the cause of chest pain. A recent study used this acid perfusion test to assess sensitivity and found a close correlation between severity of oesophagitis grade and an intensity score, which was derived from the area under the curve of sensation against time (Fig. 8). This score fell in some, but not all patients after healing of oesophagitis with proton pump inhibition. While no normal subject experienced pain during acid perfusion, balloon distension in the proximal oesophagus generated chest pain and heartburn in both patients and healthy volunteers who required similar volumes to elicit symptoms. Thus, patients with oesophagitis demonstrate sensitivity to acid perfusion only, whereas sensitivity to other modalities appears normal.⁷⁰ Other authors have come to similar conclusions. Drewes et al.⁷¹ used multimodal testing to assess the sensitivity of the oesophagus to both distension as also hot and cold stimulation and electrical activation of nerves. These authors found enhanced sensitivity to hot stimuli, normal sensitivity to cold stimulation, and sensitivity distension was, if anything, slightly reduced. The area where pain was referred was strikingly increased in patients with oesophagitis with a mean area of 49 cm², compared with 24 cm² for the healthy controls. This shows evidence of central sensitization, which is likely to occur in the spinal cord. Other authors have demonstrated that this can be readily induced even in normal individuals by infusing acid into the oesophagus.⁷² In addition to changes in innervation, there may also be important changes in receptors and immunostaining for TRPV1 was increased in individuals with oesophagitis.⁷³

Helicobacter pylori gastritis

Helicobacter pylori gastritis induces a superficial gastritis with polymorphonuclear neutrophil infiltrate in the mucosa and a chronic inflammatory infiltrate in the submucosal area, which does not extend into the muscle layer. Establishing an animal model of H. pylori gastritis has proved difficult, but Bercik et al.⁷⁴ using the Sydney strain of H. pylori were able to induce a chronic lymphocytic gastritis in BAlb/c mice. Ten months after infection, the mice showed increased mucosal and spinal cord SP and calcitonin-gene-related peptide (CGRP). These abnormalities had substantially resolved 2 months following eradication. Whether these neuropeptide changes are associated with visceral hypersensitivity is unknown. Data in humans are conflicting. Helicobacter pylori gastritis does not appear to consistently alter gastric motility or emptying. Two fairly large studies concluded that H. pylori infection did not alter visceral sensitivity.^75,76 Mearin et al. did show a numeric increase in abdominal discomfort, but this did not achieve statistical significance. Somewhat paradoxically, a much smaller study of FD, with just 13 H. pylori positive individuals, showed decreased thresholds for discomfort and an increase in CGRP and SP in the antrum. When subjects were divided according to the severity of histological inflammation, there was a striking increase in CGRP with the more severely inflamed.

Most interestingly, there was a significant negative correlation between the thresholds for discomfort induced by balloon distension and the concentrations of both CGRP and SP in mucosal biopsies. It is clear that further studies are needed and should include individuals with more severe inflammation, but this adds some support to the idea that chronic inflammation mediates hypersensitivity via changes in neuropeptides.

Intestinal inflammation and visceral hypersensitivity

The small intestine is vulnerable to infection. In human evolution, the major cause of morbidity until recent times was helminthic infestation. The response to these organisms is designed to expel them and consists of vomiting, profuse secretion with enhancement of propulsive motility. These changes are mediated in part through increased sensitivity of afferent nerves, which mediate these secretomotor effects. The detailed time course of evolution of these inflammatory events cannot realistically be studied in humans but there are several animal models, which show initial decrease followed by long-term increases in SP and other neuropeptides.^77,78

Colitis

One of the earliest studies using rectal balloon distension in 18 patients with active UC showed a decreased volume required to induce discomfort and reduced rectal compliance.⁷⁹ The same study showed that compliance was also reduced in quiescent colitis, being intermediate between active colitis and normal values.⁷⁹ Although abnormalities of neuropeptides have been reported in biopsy homogenates in chronic UC,⁸⁰ much of the SP may well be non-neuronal.

Specific immunohistochemistry is needed to identify changes in SP containing neurons, which are increased in the rectal mucosa of patients with UC, the density of staining being proportional to the severity of disease as assessed by lifetime use of steroids.⁸¹ Substance P immunoreactivity is reduced in active UC, as found in the acute lesions of a model of colitis induced by trinitrobenzenesulphonic acid (TNBS),⁸² but is increased in chronic UC together with increases in SP binding sites.⁸³ Recent studies have indicated again that an increase in inflammation in UC lowers the threshold for pain induced by rectal distension.⁸⁴ However, the same study compared the effect of repetitive sigmoid stimulation on rectal sensitivity in both IBS and UC. Irritable bowel syndrome when compared with UC patients showed a significantly reduced threshold for pain, which fell further after sigmoid stimulation. In contrast, in UC, no such fall in threshold was noted.⁸⁴ This demonstrates that chronic inflammation alone does not lead to visceral hypersensitivity. Visceral hypersensitivity depends not only on the peripheral input but also on descending influences which can be both facilitatory and inhibitory. Several other studies suggest that IBS is characterized by a defective descending nociceptive inhibitory control.^65,85 Earlier studies in Crohn's disease suggested that a chronic inflammatory disease might in fact enhance coping mechanisms. Thus, in a study of 20 patients with Crohn's ileitis, thresholds for rectal distension were increased, compared with normal and there was decreased autonomic arousal in response to distension.⁸⁶ These changes in rectal sensitivity induced by distal inflammation are likely to be mediated centrally. It may well be that coping with a chronic disease induces a degree of tolerance and enhanced coping, which IBS patients characteristically appear to lack.⁸⁶

Postinfective IBS

While most bacterial gastroenteritis is associated with mucosal ulceration in the distal small bowel and proximal colon, it has nevertheless been demonstrated that they have lower thresholds for discomfort on rectal distension.⁹ The changes in this region (distal from the area of maximal inflammation) are likely to be mediated by circulating T cells. Increased numbers of these cells are found in rectal mucosa.^9,87 These seem to mediate the EC hyperplasia characteristic of these patients, as lymphocyte and EC numbers correlate quite closely⁸⁷; and in mice lacking the T-cell receptor, EC hyperplasia induced by Trichinella infection is absent.⁸⁸ A study from China demonstrated not only increased lymphocytes but also increased mRNA for IL-1 (a macrophage product). These authors also found increased numbers of SP, 5-HT and CGRP positive nerve fibres in both terminal ileum and rectosigmoid regions, a feature of both postinfective and D-IBS without an infectious origin.⁸⁹

Whether these abnormalities relate to changes in sensitivity are unknown, although there have been recent reports that IBS patients showing visceral hypersensitivity have increased EC numbers.⁹⁰ This would be in agreement with animal studies, which indicated that, after Trichinella infection in which EC hyperplasia is seen, there is increased afferent firing in response to distension, which appears to be mediated through serotonin and is blocked by the 5-HT3 receptor antagonist ondansetron.⁹¹

Diverticulitis

Recurrent pain in diverticular disease can be a persistent problem in a small minority (1 in 10) of those with diverticulosis.⁹² A major risk factor appears to be a prior history of acute diverticulitis.⁹² Increased tachykinin staining in mucosal biopsies has been noted in symptomatic patients with diverticular disease. An abnormal pattern of neuropeptides similar to that seen in TNBS colitis (increased mucosal galanin and SP in the muscle layer) has been reported in full-thickness sections of resected colon from patients with complicated diverticular disease.⁹³ Hypersensitivity to rectal distension has been reported in diverticular disease,⁹⁴ but whether this hypersensitivity is related to neuropeptide expression remains to be established.

Conclusion

There are wide regional variations in the type and depth of inflammatory injury in the GI tract and a corresponding variability in the relationship between inflammation and sensitivity, which depends critically on the site and type of injury. Oesophagitis, which is a superficial inflammation, leads to increased expression of mucosal TRPV1 and hypersensitivity to acid and heat, but not to balloon distension. Chronic gastritis due to H. pylori also causes a superficial mucosal lesion with no clear association with hypersensitivity to distension. For reasons of inaccessibility, there are few data in humans on the relationship between small intestinal mucosal abnormalities and visceral hypersensitivity. Most data relate to the rectum. Acute inflammation of the rectum leads to decreased compliance and reduced tolerance to distension. Hypersensitivity detected using the barostat is found in ulcerative proctitis, although chronic ulceration does not appear to induce a hypersensitive state and may induce reduced reactivity in some, likely through altered central pain processing. Postinfective IBS leads to reduced compliance, but whether this is associated with true hypersensitivity is unknown. Diverticulitis and Crohn's disease are associated with deep ulceration and increased expression of a range of neuropeptides, including tachykinins and VIP.

Patients with diverticular disease also show hypersensitivity to colorectal distension. Inflammation varies widely through the GI tract and appears to alter different sensory modalities at different sites. The type of infiltrating cells and depth of inflammation may be major determinants of how high up the afferent signalling pathway visceral stimuli reach, and hence whether they reach consciousness. Superficial mucosal inflammation alone is probably insufficient to lead to visceral hypersensitivity, which is powerfully influenced by central processing. Pain associated with deeper inflammation is likely to be less influenced by central factors.

Future work will be needed to establish the precise relationship between abnormalities of neuropeptide expression and abnormalities of different modalities of visceral sensation. This knowledge should guide the design of rational treatments for postinflammatory symptoms in these conditions.

Introduction

Patients with IBS often report symptoms like diarrhoea and constipation, which suggest the existence of a disorder of intestinal motility. However, using standard techniques for investigation of intestinal motility, no specific motor disorders have been demonstrated in these patients.⁹⁵ In addition to symptoms related to altered transit of chyme, gas-related symptoms like bloating and abdominal distension are often reported by patients with IBS.^96,97 Visceral hypersensitivity could be responsible for symptoms elicited by intestinal gas in these patients. However, the coexistence of motor and sensory disturbances have been largely suspected as determinant in the origin of symptoms in functional gut disorders. This section summarizes current knowledge on the alterations of intestinal gas transit in patients with bloating, with special emphasis on the interaction between altered gas transport, altered reflex responses to visceral stimuli, and the mechanisms by which gas pooling can modulate visceral sensitivity and produce abdominal symptoms.

Intestinal gas metabolism

Every day, large amounts of gas arrive in the intestine, but most are eliminated to maintain the volume of intraluminal gas relatively low, about 100–200 mL in normal conditions.⁹⁸ The main sources of gas are swallowed air and gas produced by chemical reactions, diffusion from blood and bacterial fermentation by colonic bacteria. The main mechanisms that operate to eliminate gas are belching, diffusion of gases, bacterial consumption and transit and evacuation of gas by flatus.^98,99 Hence, gas homeostasis seems to be a complex, finely regulated process that results from the interplay between gas producing and gas consuming mechanisms (Fig. 9). A failure of these mechanisms could result in an increment in the volume of intraluminal gas and produce abdominal symptoms, such as bloating or abdominal distension.

Volume and composition of intestinal gas

Different methods have been employed in the past to measure the volume of intestinal gas. A landmark paper first described the use of a wash-out technique with a high rate of intraluminal infusion of argon.¹⁰⁰ Thus, the volume of intestinal gas was calculated to range between 100 and 200 mL.¹⁰⁰ These original results, obtained from a population in the Midwest United States, have been confirmed and shown to be similar in a Mediterranean population in the south of Europe.¹⁰¹

Intestinal gas is composed of five main gases: nitrogen, oxygen, carbon dioxide, hydrogen and methane gas.¹⁰⁰ In addition, there are traces of other gases, like sulphur gases, that are responsible for the odour of flatus.¹⁰² The concentration of intestinal gases varies depending on the nature of ingested food.^103,104 Hence, ingestion of meals rich in non-absorbable carbohydrates markedly increases the volume of hydrogen and methane gases which result from bacterial fermentation. By contrast, when the predominant gas is nitrogen, air swallowing is presumably the origin.¹⁰⁵

Sources of intestinal gas

There are four main sources of intestinal gas: swallowed air, chemical reactions, diffusion of gas and bacterial fermentation (Fig. 9). During meals, each swallow is preceded by a gas bubble of about 15 mL that enters the stomach.¹⁰⁶ Under some circumstances, for instance in patients with aerophagia, air can be swallowed by a compulsive manoeuvre and immediately belched. In some individuals, the volume of air swallowed exceeds the amount belched and the stomach becomes distended by air as it may be appreciated by a plain radiograph. The gases introduced by swallow mechanisms are nitrogen and oxygen, the main components of air.⁹⁸

At the duodenal level, acid hydrogen ions emptied from the stomach react with bicarbonate of biliary and pancreatic origin. This reaction will result in the production of carbon dioxide plus water. Several litres of carbon dioxide may be produced daily by this mechanism depending on the passage of acid from the stomach.¹⁰⁷

Gas can also diffuse from the bloodstream to the intestinal lumen. The two key factors that will determine diffusion of gases are the pressure gradient of gases between venous blood and the intestinal lumen, and the diffusion potential of each individual gas.^108,109 The gases that can diffuse by this mechanism are nitrogen, oxygen and carbon dioxide. The first has a high blood to lumen pressure gradient (partial pressure of 600 mmHg in venous blood), but a low diffusion potential. By contrast, oxygen and carbon dioxide have smaller partial pressure in venous blood, but high diffusion potential.

Non-absorbable carbohydrates are fermented by colonic bacteria, leading to the production of hydrogen and carbon dioxide within the colonic lumen. Large amounts of these gases can be produced depending on the content of the meals and the composition of colonic microflora.^110,111 About 90% of the Western population carries hydrogen-producing bacteria. In addition, about 30% of the population also has methanogenic bacteria in the left colon that consume hydrogen, releasing methane gas.^111,112

Elimination of intestinal gas

There are four main mechanisms to eliminate intestinal gas: belching, diffusion, bacterial consumption and flatus.⁹⁸ Belching is a quick and effective mechanism to release excess gas from the stomach. Recent studies using an impedance technique have shown that, in addition to gastric gas, gas of oesophageal origin can also be eliminated by the so-called supragastric belching.¹¹³ The gases mainly eliminated by this mechanism will be swallowed gases, that is, nitrogen and oxygen.

In addition to diffusion of gas from the blood to the intestine, gas can also diffuse in the opposite direction, from the intestinal lumen to the bloodstream. As previously described, diffusion will depend on the pressure gradient of gases and the diffusion capacity of each gas.^108,109 Moreover, the time of exposure of the gas to the intestinal mucosa is a key determinant of the rate of gas diffusion from lumen to blood. Hence, the interplay between gas production, transport and evacuation will define the prevailing mechanism of gas elimination from the gut. Due to the great partial pressure of nitrogen in venous blood and its poor diffusion capacity, diffusion of this gas will be difficult and slow. By contrast, diffusion is the predominant mechanism to clear carbon dioxide generated in the upper small bowel, and gases produced by colonic bacteria.^98,99

Colonic bacteria consume gas for their own metabolism. Both hydrogen and carbon dioxide are consumed in large quantities, and aerobic bacteria also consume oxygen. In addition to the aforementioned consumption of hydrogen by methanogenic bacteria, most subjects have a pool of sulphate-reducing bacteria in the colon that metabolize hydrogen and release small amounts of sulphur-containing gases that are responsible for the characteristic odour of flatus.^111,112,114

Intestinal gas is transported in an oro-caudad direction and expelled by flatus. Although the exact motor events that are involved in transit and evacuation of gas are, by and large, unknown, flatus emission has been associated with peristaltic activity in the distal colon and with relaxation of the anal sphincter.¹¹⁵

Transit and evacuation of intestinal gas

Under basal circumstances, the human intestine is able to propel and evacuate large gas loads quickly and effectively, preventing gas retention and abdominal symptoms¹¹⁶ (Fig. 10). The motor events implicated in gas transport and evacuation are incompletely known. Previous studies have failed to identify any form of phasic activity that could be plausibly linked to gas transit.¹¹⁷ However, it has been shown that gas infusion into the jejunum elicits a sustained tonic contraction proximal to the gas infusion site, which suggests that changes in tonic motor activity may play a role in displacement of intraluminal gas.¹¹⁸ Likewise, inhibition of motor activity by intravenous administration of glucagon disrupts gas transit and evacuation, indicating that transit of gas is an active process that requires intestinal motor activity, rather than passive displacement of low resistance gas masses.¹¹⁹

Under normal conditions, gas transit and evacuation are modulated by several factors including intraluminal factors, such as mechanoreceptor activation by gut distension and chemoreceptor activation by nutrients, and extra-abdominal factors like physical exercise and body posture. Gut distension is an important trigger of gas movement that seems to be operative in various segments of the intestine like the stomach, small bowel and rectum.^120,121 The intensity of distension required to generate gas motion is generally mild and well tolerated, and even unperceived distension can promote transit of gas, a phenomenon that is common with most intestinal reflexes that may be activated by stimuli below the threshold for perception.

Nutrients modulate transit of intestinal gas, depending on their composition, caloric load and site of action. Lipids and proteins, but not carbohydrates, have been shown to slow down gas transit and evacuation when infused into the proximal duodenum.¹²² Duodenal lipids act in a dose-dependent manner to inhibit the transit of gas. Thus, caloric loads of a minimum of 1 kcal min⁻¹ are required to produce measurable changes in gas transit in health.¹²³ This concentration is in the range of the 1–2 kcal min⁻¹ of similar lipid mixtures usually used to inhibit intestinal transit of chyme through the small intestine. However, much smaller caloric loads (about 0.3 kcal min⁻¹) infused into the ileum have been shown to delay gastric emptying.¹²⁴ Likewise, lipids infused into the ileum have a much greater inhibitory effect on gas transit and evacuation, which can lead to significant retention of intestinal gas.¹²⁵ The clinical relevance of this phenomenon in patients with accelerated intestinal transit (as in diarrhoea) and bloating deserves further study.

Physical activity and posture have also been shown to modulate intestinal transit of gas. Hence, transit and evacuation of gas is quicker in the upright than in supine position.¹²⁶ Likewise, mild physical activity has been shown to accelerate transit and evacuation of intestinal gas, an effect that is achieved at levels of activity that do not affect heart rate or blood pressure.¹²⁷

Intestinal transit of gas in patients with bloating

Using gas infusion techniques, it has been shown that patients with IBS and abdominal bloating have an altered transit of intestinal gas.^101,123,128 Abnormal gas transit was observed both when caudal gas recovery was performed using an external intergluteal cannula that did not interfere with anorectal function and when gas was recovered via an internal rectal cannula that avoided any potential effect of the pelvic floor or the anus on the ability to evacuate the gas. Hence, the site for altered gas transit in IBS must be located at intestinal or colonic levels proximal to the anorectal level. Further studies using scintigraphic imaging with radioactive 133-xenon have, in fact, shown that the small bowel is the main intestinal segment responsible for delayed transit of gas in these patients.¹²⁹ Likewise, the small bowel appears to be responsible for the retention of gas observed in patients during lipid infusion because, when gas is infused during concomitant lipids in patients with IBS, gas transit is also delayed through the small bowel, with no additional recruitment of distal segments of the gut.¹³⁰

The type of motor activity responsible for altered transit in IBS patients is unknown, but it may be normalized by cholinomimetics, suggesting that a decrease in motor activity, rather than an increased activity with functional obstacle to flow is involved in the delayed transit and retention in patients with bloating.¹²⁸

We have also shown that altered transit of intestinal gas in patients with significant bloating symptoms is associated with an altered reflex response in the small bowel, such as the normal response to intestinal distension.¹³¹ Moreover, patients with IBS are hypersensitive to the braking effect of lipids on gas transit so that small concentrations (0.5 kcal min⁻¹) of lipids that exert no effect on gas transit in healthy subjects are great enough to delay intestinal transit of gas in these patients.¹²³ Hence, alteration in the reflex mechanisms that normally modulate intestinal transit of gas may contribute to gas retention and symptoms in patients with IBS and/or functional bloating.

Intestinal gas and abdominal symptoms

Under certain conditions, intestinal gas can produce two main types of abdominal symptoms: symptoms due to excessive gas evacuation (belching and flatulence) and symptoms due to excessive gas retention (bloating and pain).

Belching is a normal mechanism to evacuate excess gas from the stomach. Patients who complain of excessive belching often have associated aerophagia, that is, the patient swallows air that is finally belched, leading the patient to erroneously conclude that there is excessive production of gas in the stomach.^98,99

Flatulence may be related to excessive gas evacuation or to annoying evacuation of odoriferous gases, which are sulphur-containing gases generated by colonic bacteria.¹⁰² An increment in the volume of gas evacuated could be due to swallowed air or to impaired gas absorption, but the mechanism commonly responsible for excessive gas production is fermentation of undigested substrates by colonic bacteria.¹⁰⁵ Healthy asymptomatic subjects pass gas about 10 times per day, but there are great variations in the volume of gas evacuated depending on the composition of meal.¹³² There are few cases of documented true increments of gas evacuation in flatulent patients.¹⁰⁵ Some studies have suggested that patients with IBS may have excessive gas production linked to excessive fermentation and bacterial overgrowth¹³³ but, in general, there has been a discrepancy between documented carbohydrate malabsorption and abdominal symptoms reported in the literature.^134,135

Bloating is common among patients with IBS.^96,97 Bloating is often considered to be related to retention of intestinal gas, and is relieved by flatus or belching. As previously described, patients with abdominal bloating cannot propel and evacuate gas loads, a failure that leads to retention of gas. Gas retention in these patients is associated with perception of abdominal symptoms (Fig. 11). Interestingly, these symptoms, mainly bloating and borborygmi, are often recognized by patients as their customary complaints.^101,123,128 Hence, gas infusion in patients complaining of bloating produces gas retention and reproduces patients’ usual symptoms. However, the correlation between the volume of gas retained by patients and the intensity of perception referred is poor (Fig. 11), suggesting that factors other than gas volume influence the perception of symptoms.^101,123

Visceral hypersensitivity is highly prevalent in patients with functional gut disorders and can play a key role in symptom generation.^51,136 Hence, in addition to the widespread evidence of visceral hypersensitivity in these patients, we have also observed that at similar volumes of intestinal gas retention, patients with abdominal bloating develop significantly greater abdominal symptoms than healthy controls despite similar distribution of gas along the intestine.¹⁴²

Gas distribution along the intestine may determine the final perception of symptoms. In healthy subjects, retention of a certain volume of gas distributed along the small bowel and the colon produced significantly greater symptoms than retention of the same volume of gas pooled in the distal colon.¹¹⁸ Spatial summation of stimuli is a phenomenon that may be responsible for differences in symptom intensity according to gas distribution along the gut. Hence, pooling of gas in long segments of the intestine will not only produce greater perception, but also sensitize other segments of the intestine, and thereby could, theoretically, be mechanistic in the induction of visceral hypersensitivity in patients with gas retention.^137,138

Motor activity of the intestine also influences perception of gas retention. Hence, in a model of induced gas retention, subjects developed significantly greater perception during obstructed gas evacuation with preserved intestinal motility than when intestinal motility was pharmacologically abolished by glucagon.¹¹⁹

Objective abdominal distension

The sensations of abdominal bloating and abdominal distension may be linked, but not invariably so. Reliable methods to measure abdominal distension and the variations in abdominal circumference during the day have been described, but they have methodological limitations. For instance, previous studies using tape measures of abdominal girth have provided variable results.^139,140 More recently, an ambulatory technique using inductance plethysmography has shown greater variations in girth in patients complaining of bloating than in healthy subjects. However, the authors found a good correlation between distension and subjective bloating only in patients with constipation-predominant IBS (C-IBS), but not in D-IBS.¹⁴¹

Abdominal distension is determined by the correlation between the volume of intraabdominal contents and the muscular activity of the abdominal wall. In our laboratory, we have shown a good correlation between abdominal distension and the volume of gas retention.¹⁰¹ Thus, in a study using a model of induced gas retention in healthy subjects, the increment in abdominal girth was the same whether gas was widely distributed along the intestine or confined to the distal colon.¹¹⁸ However, in patients with abdominal bloating, abdominal distension was greater than in healthy subjects at similar intraluminal gas volumes.

This finding may be explained by the important differences that were observed between patients and healthy subjects when the muscular activity of the abdominal wall was studied by electromyography. Whereas healthy subjects contracted the abdominal muscles in response to increments in intestinal gas volume, patients failed to do so and even manifested a paradoxical relaxation of certain abdominal wall muscles like the internal oblique, which is the muscle normally activated to support the abdomen in the upright position. Hence, altered viscerosomatic responses to increments in intraabdominal contents like gas probably contribute to abdominal distension in patients with bloating.¹⁴²

Conclusion

Patients with functional gut disorders often refer symptoms that they attribute to excessive intestinal gas. In recent years, the development of new methodologies for the study of different aspects of gas physiology has increased our understanding of the physiology of intestinal gas and provided new insights into the pathophysiology of functional gut disorders. Patients with abdominal bloating appear to have impaired ability to caudally transport and evacuate intestinal gas loads. This pathophysiological abnormality is associated with an alteration of reflex mechanisms that normally modulate intestinal gas transit. The volume of gas retained in the gut will determine the magnitude of the objective abdominal distension often associated with bloating. By contrast, perception is determined by other factors such as the distribution of gas in the different segments of the intestine and intestinal motor activity. In addition, patients with bloating manifest altered viscerosomatic reflex responses to intestinal volume increments that may also contribute to the development of abdominal distension. This new information, when considered together with the well-documented phenomenon of visceral hypersensitivity, leads to the hypothesis that patients with IBS and functional bloating suffer from altered processing of intraluminal signalling that produces, both an exaggerated perception of intestinal stimuli at central levels, and also altered peripheral processing of intraluminal stimuli and impaired adaptive reflex responses (Fig. 12). In some individuals, these pathological effects may be compounded by paradoxical relaxation of the internal oblique muscles of the abdomen, with a hanging distended abdomen in the upright position. The interplay of abnormal responses to intestinal stimuli may determine the prevalent symptom pattern (pain, bloating and distension) in each individual patient.

Introduction

While visceral hypersensitivity is perceived to occur in about 40–60% of patients with FGIDs, there is evidence supporting a role for other peripheral mechanisms in the pathobiology of these disorders. Some of these factors may be under genetic control.

The objectives are to briefly review the contribution of bacteriological mechanisms, non-inflammatory intraluminal factors and genetics to IBS. An important molecule at the interface between intraluminal milieu and the human organism is serotonin, which is located in the enteroendocrine cells in the lining of the GI tract and serves as a mediator of chemical and mechanical transduction of luminal stimuli.

Serotonin synthesis and reuptake in the gastrointestinal epithelial lining

Aficionados have claimed that IBS is a disorder of serotonin (5-HT). This is partly influenced by the recent evidence that modulation of serotonergic mechanisms significantly impact the manifestations of IBS. There are direct lines of evidence to address the potential role of serotonin. A detailed account of the role of serotonin in mediating sensory, secretory and motor responses in the gut is beyond the scope of this review. However, insights from observations in patients with IBS will be briefly discussed, as this represents a model for the potential of several other neurotransmitters and bioactive molecules to transduce the effects of the intraluminal milieu on the biology of the GI lining, e.g. cholecystokinin, peptide YY, SP and neuropeptide Y.

Plasma 5-HT levels are elevated in patients with IBS.^143,144 Postmeal symptoms can be prominent in IBS patients,^144,145 and Houghton et al. attempted to correlate postprandial symptoms and increase in plasma 5-HT.¹⁴⁴ However, it appears that the peak in plasma 5-HT levels does not match with the time (first 60–90 min) when patients typically experience pain, diarrhoea and urgency in IBS.¹⁴⁵

Enteroendocrine cells in rectal biopsies of postinfectious IBS patients^11,87,89 show significantly increased numbers of 5-HT-containing cells. These quantitative differences were more impressive in one study¹¹ than a subsequent study from Spiller's group⁸⁷ and overlap with numbers in disease control groups or non-postinfectious IBS, suggesting that prior infection may not be a key factor.

Several factors involved in the control of 5-HT in IBS were assessed using mucosal biopsies from IBS, healthy and disease controls (UC). Thus, Coates et al. showed that the number of EC cells containing 5-HT was normal,¹⁴⁶ in contrast to other studies,^87,89,147 and the release of 5-HT from mucosal biopsies under baseline conditions or in response to stimulation was normal. However, the mucosal content of the 5-HT reuptake protein, SERT, was reduced as shown by SERT mRNA content and immunohistochemistry.¹⁴⁶ These changes paralleled the findings in non-severe UC, and it is unlikely that the diarrhoea itself may have induced the changes, as results were similar in D-IBS and C-IBS. The mechanisms causing these changes and their functional importance remain the subject of continued research. On the contrary, a study performed at Mayo Clinic¹⁴⁸ did not confirm the findings of Coates et al.¹⁴⁶. Andrews et al. performed a study of both rectal and sigmoid mucosa and were unable to show any significant differences in the SERT mRNA content between 15 patients with D-IBS, 15 with C-IBS and 15 healthy controls.¹⁴⁸ Careful review of biopsies taken from adjacent sections of macroscopically normal mucosa revealed no differences in lymphocytic infiltration when scored by a specialist GI histopathologist who was blinded to information on the clinical status of the participants whose biopsies were coded.

An additional precaution in the Mayo study is that a separate ‘confirmatory’ cohort of healthy controls and patients with IBS was also assessed and the findings were confirmed. An intriguing difference in studies by Coates et al. and Andrews et al. is that the former found a ratio of 10⁻⁵ SERT mRNA to β-actin, whereas the latter found median SERT mRNA to control proteins (including β-actin) ratio of 0.5. The reasons for these quantitative differences in the two studies are unclear.

Several groups have explored genetic epidemiology of serotonergic control in IBS, particularly the potential role of SERT-P genotype on the presentation of IBS.^149–152 There are large variations in the distribution of different genotypes in controls of different ethnicities. Hence, comparisons must include ethnically matched groups. All studies of IBS, overall or of C-IBS, appear to be unrelated to SERT-P genotype. Two studies show a significant association of SERT-P genotype with D-IBS. In Turkey, a study of 18 D-IBS patients showed a significantly lower proportion of patients with homozygous long and increased heterozygous genotype.¹⁴⁹ In a multicentre study in the United States, compared with commercially derived control data, there was a relative increase in the proportion with homozygous short genotype, although interestingly, there was an equal split of the three genotypes in this cohort.¹⁵⁰. In the largest single-centre study to date, with 128 D-IBS and 120 controls, there was no significant odds ratio for any genotype with D-IBS.¹⁵¹

Differences in SERT function appear to influence the response to therapy in IBS,¹⁵³ suggesting there is an important pharmacogenomic effect on SERT biosynthesis; further studies are awaited.

Serotonin is a key player in the secretion and dysmotility of carcinoid syndrome,¹⁵⁴ in which 5-HT₃ antagonists reduce the colonic response to feeding,¹⁵⁵ diarrhoea, and the need-for-rescue antidiarrhoeal agents.¹⁵⁶ In summary, the concept that serotonin is a potential major player in IBS is compelling, and the response to serotonergic agents in clinical trials in IBS¹⁵⁷ is also supportive. Further studies of other transmitters released from the GI epithelium may identify other targets for treatment in IBS. Release of these bioactive substances may result from changes in the intraluminal milieu.

‘Irritated bowel syndrome’: the intraluminal milieu

In this section, the concept of the irritated bowel will be addressed because recent data suggest that changes in the intraluminal milieu may result in symptoms, which may be restored by treatments directed at the intraluminal factors.

There is overlap in symptoms between idiopathic bile acid catharsis or rapid ileal transit that results in failure of bile acid absorption. Williams et al.¹⁵⁸ used bile salt retention to demonstrate that functional diarrhoea may or may not be associated with pain and that such patients may have evidence of bile acid malabsorption. Patients with IBS show higher intestinal secretion in response to perfused bile acids in the ileum relative to controls.¹⁵⁹ Sciarretta et al. showed that retarding transit with loperamide reduced the risk of bile salt loss.¹⁶⁰ It is not surprising, therefore, that bile salt binding with cholestyramine may be effective in the treatment of IBS with diarrhoea.

Perfusion of the mammalian or human colon with di-α-hydroxybile acids such as chenodeoxycholic acid or short-chain fatty acids (SCFAs) results in high-amplitude propagated contractions or rapid transit in healthy colon.^161–165 Studies have not been conducted in IBS patients. However, the relatively low concentrations of bile acid (1 mmol L⁻¹) required to induce highly propulsive propagated sequences suggest that relatively modest levels of malabsorption may be sufficient to perturb colonic motility. This has also been demonstrated in rodents, suggesting that this is a conserved response in many species.¹⁶⁶ Short-chain fatty acids result in increased 5-HT release from the mucosa in rodents.¹⁶⁶

There are experimental data that suggest SCFAs decrease the ratio of non-propulsive to propulsive activity via PYY.¹⁶⁷ It is intriguing that some of the biopsies from patients with postinfectious IBS actually show increased PYY containing cells in the mucosa. It may be postulated that the increased PYY serves to mediate the effect of intraluminal SCFAs.

In summary, nutrients, products of digestion and endogenous factors may release bioactive substances that trigger motor and sensory changes and may contribute to visceral hypersensitivity in IBS.

Bacterial ecology in the intestine of patients with IBS

Small studies have documented differences in bacterial populations in IBS subgroups.¹⁶⁸ An intriguing study suggests that there are significantly different bacterial counts in IBS patients according to predominant bowel dysfunction: decreased Lactobacillus species in D-IBS, increased Veillonella species in C-IBS.¹⁶⁹ These results are worthy of further study. This is also intriguing, given the observation by Husebye et al. that intestinal motor migrating complex period in experimental animals changes after introduction of specific intestinal microflora.¹⁷⁰

The role of bacteria in bowel function and symptoms may also be mediated through gas production. The presence of methane in lactulose breath test among IBS subjects was highly associated with the constipation-predominant form.¹⁶⁸ Therefore, the authors set out to determine whether methane gas can alter small intestinal motor function in dogs with small intestinal fistulae created to measure intestinal transit of a radiolabel. Transit was evaluated during infusion of room air and subsequently methane, which slowed transit in all dogs by an average of 59%. In a second experiment, guinea pig ileum was pinned into an organ bath for the study of force of contractile activity orad and aborad in response to brush strokes. Gassing the bath with methane augmented contractile activity orad and aborad to the stimulus, compared with room air. In the third experiment, humans with IBS having undergone a small bowel motility study were compared such that subjects producing methane on lactulose breath test were compared with those producing hydrogen. The small intestinal motility index was significantly higher in methane-producing IBS patients, compared with hydrogen producers. Therefore, methane, a gaseous by-product of intestinal bacteria, slows small intestinal transit and appears to do so by augmenting small bowel contractile activity.

Thus, changes in motor function may result from bacterial factors in the gut. Little is known about the interaction of specific bacterial species and development of IBS symptoms; however, the preceding section demonstrated the potential for excess gas volume or gas transit to evoke abdominal symptoms irrespective of visceral hypersensitivity.

Effect of prior infection on intraluminal bacteria, persistent inflammation and development of FGID

Epidemiological studies suggest development of dyspepsia and IBS during a 1-year follow-up in a cohort of adult patients affected by a Salmonella enteritidis acute gastroenteritis outbreak.¹⁰ Before the acute gastroenteritis outbreak, the prevalence of dyspepsia was similar in cases and controls (2.5%vs 3.8%); the prevalence of IBS was also similar (2.9%vs 2.3%). Salmonella gastroenteritis is a significant risk factor not only for IBS but also for dyspepsia; at 1 year of follow-up, 1 in 7 and 1 in 10 subjects developed dyspepsia or IBS, respectively.

Similarly, a waterborne infection in Ontario, Canada resulted in increased IBS prevalence relative to uninfected controls.¹⁷¹

Other studies suggest that probiotics or antibiotics aimed at changing bacterial counts may indeed alter symptoms. Although the mechanism is still incompletely understood,^172–174 one important mechanism is a restoration by a probiotic Bifidobacterium species of normal IL₁₀ to IL₁₂, suggesting that a pro-inflammatory state may have been the result of an antecedent inflammation or infection. Similarly, other probiotics independently shown to provide symptomatic benefit in patients with IBS (e.g. on bloating and flatulence) have known effects on inflammatory mediators in animal models, as reviewed elsewhere.¹⁷⁵

Nobaek et al. showed that alteration of intestinal microflora by feeding probiotics was associated with reduction in abdominal bloating and pain in patients with IBS.¹⁷² Similarly, Kim et al. showed that, in patients with IBS, the combination probiotic, VSL#3, resulted in improvement of the symptom of bloating, but no other symptoms, global relief, or change in small bowel or colonic transit were observed relative to placebo treatment.¹⁷⁵ Quigley et al. have preliminarily reported that there is benefit with Bifidobacteria, but not with Lactobacillus-containing probiotics.¹⁷⁴

The probiotic, Lactobacillus GG, was evaluated under randomized, double-blinded, placebo-controlled conditions for Rome II IBS in 50 children for 6 weeks. Lactobacillus GG was not superior to placebo in relieving abdominal pain (40.0% response rate in the placebo group vs 44.0% in the Lactobacillus GG group) or other GI symptoms except for a lower incidence of perceived abdominal distension favouring Lactobacillus GG.¹⁷⁶ These effects on distension are consistent with the reduced bloating and flatulence reported in the two Mayo Clinic studies of VSL#3.^173,175

In summary, a subgroup of IBS is associated with prior infection and may cause subtle inflammation. Probiotics appear to restore local immune function and, in one study, slowed colonic transit in D-IBS. The role of probiotics in visceral hypersensitivity is unclear.

Small intestinal bacterial overgrowth in IBS

Recent reports suggest bacterial overgrowth is commonly associated with IBS when diagnosed using the lactulose–hydrogen breath test. One group assessed by means of a case–control study the prevalence of small intestinal bacterial overgrowth by 50 g glucose–hydrogen breath test (positive by peak of H₂ values >10 ppm) in 65 patients with Rome-II-criteria-positive IBS symptoms, compared with a control group of 102 sex- and age-matched healthy subjects without IBS symptoms.¹⁷⁷ Positive glucose breath test was found in 31% of IBS patients and 4% in the control group, suggesting an association between IBS and small intestinal bacterial overgrowth.

On the contrary, Walters and Vanner¹⁷⁸ employed this test to examine whether similar findings exist in a geographically distinct population of Rome II positive-IBS patients and compared 10 g lactulose breath test to the ¹⁴C-D-xylose breath test, a test with acknowledged greater specificity for bacterial overgrowth. The IBS patients were predominantly female (64%) with severe symptoms (80%), 63% had diarrhoea-predominant symptoms, and only 3% were constipation predominant. Only 10% of patients had a positive lactulose breath test, while 13% had a positive ¹⁴C-d-xylose test. In a second series of patients, no differences were found between IBS patients and controls. The authors concluded that lactulose breath test did not reliably detect a common association between bacterial overgrowth and IBS in their patient population. Further studies are clearly required.

Change in the bacterial ecology with the non-absorbable antibiotic, neomycin, resulted in short-term improvement in composite score for IBS and bowel dysfunction.¹⁷⁹ These effects were observed in a 7-day trial. Interestingly, patients whose lactulose–hydrogen breath test normalized had greater improvement than those with no response in breath hydrogen. The authors’ interpretation is that this represents an effect on bacterial overgrowth in the small bowel in IBS. While this interpretation may be questioned and longer term studies and outcomes of treatment with bacterial modification need to be evaluated, these data are consistent with the concept that it is time to revisit the role of the intestinal milieu and the effects of bacteria and endogenous factors such as bile acids and fatty acids in the mechanisms and treatment of IBS.

Similarly, antibacterial approaches, as with rifaximin, are effective in the treatment of bloating and flatulence. A randomized, double-blind, placebo-controlled trial consisting of three 10-day phases was conducted: baseline (phase 1), treatment with rifaximin 400 mg b.i.d. or placebo (phase 2), and post-treatment period (phase 3). One hundred twenty-four patients were enrolled (63 rifaximin and 61 placebo). Baseline characteristics were comparable and none had an abnormal baseline lactulose–hydrogen breath test. At the end of phase 2, there was a significant difference in global symptom relief with rifaximin vs placebo (41.3%vs 22.9%). This improvement was maintained at the end of phase 3 (28.6%vs 11.5%). Mean cumulative and bloating-specific scores dropped significantly in the rifaximin group. Among patients with IBS, a favourable response to rifaximin was noted (40.5%vs 18.2%), persisting to the end of post-treatment phase (27%vs 9.1%). H₂-breath excretion after lactulose dropped significantly among rifaximin responders and correlated with improvement in bloating and overall symptom scores. No adverse events were reported.¹⁸⁰

In summary, antibacterials reduce overall symptoms, especially bloating and flatulence, in IBS, similar to effects reported with probiotics. It is unclear whether this beneficial effect stems from a change in gas production or perturbation of visceral sensory mechanisms to reduce perception of the distension from gas.

Preliminary evidence of the role of genetics in IBS

Three lines of evidence suggest there may be a role for genetics in IBS, but the data are certainly not conclusive at present.

Firstly, familial aggregation studies suggest that family members of IBS patients are more likely to suffer from IBS than their spouse controls.^181,182

Secondly, twin studies also document a difference between risk of IBS in monozygotic relative to dizygotic twins;^183,184 however, the fact that mothers of mono- and dizygotic twins have similar prevalence of IBS suggests that heredity and social learning or other environmental factors interact as risk factors in development of IBS.¹⁸⁴

Thirdly, genetic epidemiology studies start to provide some evidence of a genetic association in IBS. However, these data are to be viewed as preliminary, and the pitfalls of such association studies, which may be underpowered should be kept in mind in reviewing these data.

Fewer IBS patients have high IL-10 producer (G/G) genotype than controls;¹⁸⁵ the potential functional significance is that these patients do not down-regulate inflammation, which may be a factor in the development of IBS. Holtmann et al. have shown that a polymorphism in the gene for the G protein involved in mediating the effects of several neurotransmitters and hormones, GNβ3 C825T genotype, is significantly associated with the report of dyspepsia and to a lesser extent in IBS patients presenting to a clinic in Germany.⁴⁰ In a study of about 250 patients with IBS with different phenotypes including overlap with uninvestigated dyspepsia, there was no evidence of a significant association between lower FGIDs or IBS with dyspepsia and GNβ3 C825T genotype.¹⁸⁶

We investigated adrenergic and serotonergic genotypes in 276 IBS patients and 120 community controls. A 44-bp deletion in the gene for the promoter for SERT (SERT-P) has been previously shown to influence the function of the SERT protein produced. Thus, the wild type l/l polymorphism results in normal function, whereas the presence of the short allele (s/l or s/s) results in impaired function.¹⁸⁷ In 90 patients with C-IBS out of 274 IBS patients and 120 controls, Kim et al.¹⁵¹ showed a significant association with α_2A-adrenoceptor polymorphism, and the combination of the α₂-adrenoceptor and SERT-P polymorphism was associated with high somatic scores in patients with lower FGIDs (Fig. 13). The SERT-P alone was not a risk factor for D-IBS in the Mayo Clinic study,¹⁵¹ and a summary of studies to date suggests there is no significant association with IBS or its subtypes^{150–152,158} in studies from North America and Korea. A positive association with D-IBS in Turkey¹⁴⁹ needs to be viewed with caution, given the ethnic differences in the genetic distribution of the polymorphism and the small sample studied (n = 18). The reduced mucosal SERT in C-IBS and D-IBS observed by Coates et al.¹⁴⁶ does not appear to be genetically determined, based on reports in the literature to date. Moreover, Andrews et al. showed only minor differences of SERT mRNA in D-IBS and not in C-IBS, and the SERT-P genotype was not a significant determinant of expression of SERT mRNA in rectal or sigmoid mucosa.¹⁴⁸

In summary, the role of genetic factors in the development of visceral hypersensitivity or the FGIDs requires further study.

Conclusions

While there is increasing evidence that IBS represents an enteric neurological disorder that alters bowel sensation and motility, it is important to continue to study other factors such as the intraluminal factors that may result in dysregulation of those functions as also the local immune system and potentially leading to inflammation. Responses to interventions with antibiotics and probiotics lend support to the importance of these mechanisms. At this time, the role of genetics beyond the clinical observation and epidemiological studies suggesting familial aggregation is uncertain. What is nevertheless clear is that the gut itself is likely to be the source of disturbed functions although the central mechanisms may alter the perception of gut sensations, imparting an unpleasant connotation to those sensations. The gut is certainly not an innocent bystander, but a potentially important target for therapeutic intervention.