Immune and mast cell activation

Low-grade inflammation (mainly characterised by a dense mast cell infiltrate) is present in colonic mucosal biopsies of about two-thirds of patients with IBS.12 Mast cells are known to release a variety of mediators, including serine proteases, which evoke neuronal hyperexcitability, a major factor for functional symptom generation (eg, pain).13 ,14 Food components, particularly proteins, may be pathogenically involved with this process, either primarily or secondarily. One possible interpretation of such mucosal changes is that food components/antigens pass through a leaky (ie, more permeable) epithelial barrier, leading to mast cell infiltration and activation, thereby leading to IBS symptoms.5 ,8 Mast cells can be activated by allergy-like mechanisms, such as those involving food-specific immunoglobulin E (IgE). However, tests for food allergy detection that use the systemic immune compartment, such as skin-prick tests, have a poor sensitivity and specificity.15 Thus, immune response to food in IBS may require more sophisticated approaches to be demonstrated.

One method is to present the offending protein to the gut immune compartment. A sort of ‘mucosal prick test’ renamed as colonoscopic allergen provocation (COLAP) test involves colonoscopy-guided submucosal injection to unravel food hypersensitivity.16 Seventy-seven per cent of a population with gut symptoms thought possibly related to food hypersensitivity had a positive COLAP test, which was consistently negative in the few control subjects. A more refined technique (confocal laser endomicroscopy) demonstrated that submucosal injection of food antigens caused increased infiltration with intraepithelial lymphocytes (IEL), formation of epithelial leaks/gaps and widening of intervillous spaces in more than half of IBS, and not in a small group of controls.17 These changes occurred within a few minutes of food antigen injection and predicted the clinical response to specific food withdrawal. Alternatively, circulating basophils have been used to determine allergens in vitro without the need to risk an allergic reaction when the patient is challenged. Indeed, basophil activation when exposed in vitro to dietary proteins, especially of wheat and milk origin, correlated with clinical responsiveness to dietary restriction of the relevant protein by one group,18 but another could find no specificity for basophil activation.19 Overall, these studies do suggest that reaction to food, whether it be via allergic, other immune or epithelial-damaging mechanisms, may play a role in the genesis of symptoms in some patients presenting with IBS. Confirmatory studies are needed before reaching diagnostic relevance.

Mechanoreceptor activation

Many different foods can evoke intestinal (luminal) distension, which, in the presence of visceral hypersensitivity and abnormal gut motility, may trigger bloating, abdominal pain and changes in bowel habit. Dietary FODMAPs are believed to act via luminal distension as discussed later.20–22

Chemosensory activation by bioactive molecules

A wide array of foods contains potentially bioactive chemicals, such as salicylates, amines, benzoates and glutamate, which can elicit neurally and/or mast cell-mediated mechanisms contributing to IBS symptoms.11 ,23 ,24 However, a specific cause–effect mechanism between bioactive chemicals and symptoms is still far from being established. Empirical clinical experience indicates an improvement of IBS symptoms as a result of reduction of the dietary intake of bioactive chemicals. Nonetheless, clinicians should have a cautionary approach before advising dietary restrictions as nutritional defects may became a serious issue for the patient.

The controversy over nomenclature

If gluten is a major trigger for IBS, it expands the gluten-related disorders by adding a new entity now referred to as non-coeliac gluten sensitivity (NCGS).27 Indeed, coeliac disease-like abnormalities were reported in a subgroup of patients with IBS many years ago.28 A recent expert group of researchers reached unanimous consensus attesting the existence of a syndrome triggered by gluten ingestion.29 This syndrome recognises a wide spectrum of symptoms and manifestations including an IBS-like phenotype, along with an extra-intestinal phenotype, that is, malaise, fatigue, headache, numbness, mental confusion (‘brain fog’), anxiety, sleep abnormalities, fibromyalgia-like symptoms and skin rash. In addition, other possible clinical features include gastro-oesophageal reflux disease, aphthous stomatitis, anaemia, depression, asthma and rhinitis. Symptoms or other manifestations occur shortly after gluten consumption and disappear or recur in a few hours (or days) after gluten withdrawal or challenge. A fundamental prerequisite for suspecting NCGS is to rule out all the established gluten/wheat disorders, comprising coeliac disease (CD), gluten ataxia, dermatitis herpetiformis and wheat allergy. The major issue not addressed by the consensus opinion was that gluten is only one protein contained within wheat. Other proteins, such as amylase-trypsin inhibitors (ATIs),30 are strong activators of innate immune responses in monocytes, macrophages and dendritic cells. Furthermore, wheat germ agglutinin, which has epithelial-damaging and immune effects at very low doses at least in vitro, might also contribute to both intestinal and extraintestinal manifestations of NCGS.31

Consequently, a further development of this research field led to suggestions of a broader term, non-coeliac wheat sensitivity (NCWS). The problems with this term are twofold. First, rye and barley may be inappropriately excluded. Second, the term will refer to any wheat component that might be causally related to induction of symptoms and, therefore, will also include fructans (FODMAPs). It will then have a very non-specific connotation in IBS. A more correct term would then be non-coeliac wheat protein sensitivity (NCWPS) since this does not attribute effects to gluten without evidence of such specificity, eliminates the issue of fructan-induced symptoms and avoids the unknown contribution of rye and barley proteins to the symptoms. Both NCGS, the currently accepted term, and NCWPS will be used subsequently in this paper.

Evidence for involvement of sensitivity to wheat proteins in IBS

Due to the lack of biomarkers, the diagnosis of NCGS still challenges clinicians as it remains based only on clinical criteria.32 ,33 In addition, one of the major diagnostic criteria for NCGS, which is the improvement of symptoms after wheat protein/gluten exclusion, might be influenced by a placebo effect experienced by patients after food elimination from their usual diet.34 This is compounded by a huge media drive, via publications, printed media, television and the internet supported strongly by celebrity endorsement where a gluten-free diet (GFD) has been embraced not only as a solution to many symptoms but also with the erroneous belief that it is healthy not to eat gluten and, even more, that GFD helps to lose weight. As a result, a high proportion of US population, for example, switched to GFD with a marked increase of the global sale of gluten-free foods.35 Because of these facts, it has been hypothesised that NCGS might be a false problem created by media rather than an emerging clinical entity.36 However, recent studies have provided strong signals that wheat protein/gluten may specifically induce GI and extraintestinal, including psychological, symptoms in at least some patients with NCGS.37–39

Although epidemiological data are still scanty and approximate, NCGS may be at least as common as CD (ie, occurring in ≥1% of the general population).40 Similarly to IBS, NCGS affects more young (third decade of life) women (F:M ratio >3:1), while, in contrast to IBS, NCGS is diagnosed more commonly in tertiary than primary care centres. According to the National Health and Nutrition Examination Survey, a primary care programme, NCGS was found in 0.6% over 7762 subjects,41 whereas at the Celiac Disease Center (University of Maryland) 6% over 5896 subjects were identified as NCGS.42 In an Italian multicentre prospective survey carried out on 38 referral paediatric and adult centres for the diagnosis of gluten-related disorders, NCGS and CD were respectively diagnosed in 391 (3.2%) and 340 (2.8%) over 12 255 patients consecutively studied in a 1-year period.40 Such data have to be viewed, however, in the light of the failure to actually prove specific sensitivity to gluten/wheat proteins in double-blind placebo-controlled (DBPC) cross-over studies in most patients fulfilling the criteria for NCGS.39 ,43

Although still a matter of debate, several factors have been postulated to play a role in NCGS pathogenesis. First, NCGS may be an immune-mediated disorder evoked by innate immunity, as highlighted by the increased expression of toll-like-receptors (TLRs), mainly TLR2, in the intestinal mucosa.44 More recently, however, the evidence of an increased level of interferon-γ in small intestinal biopsies of patients with NCGS after a short (3-day) gluten challenge lends support to a possible role exerted by adaptive immunity in this syndrome.45 In line with the latter findings, the detection of antigliadin antibodies in >50% of patients with NCGS provides further support to adaptive immunity in NCGS pathogenesis.46 Second, discordant data exist on epithelial barrier dysfunction. Initial studies showed a reduced intestinal permeability in NCGS, thus suggesting an increased intestinal barrier function. This finding has been also supported by a significantly higher expression of claudin-4 mRNA, a marker of reduced permeability, in duodenal biopsies of patients with NCGS.44 However, more recently, some evidence for increased intestinal permeability in a subgroup of patients with IBS-D carrying the human leucocyte antigen (HLA)-DQ2+/DQ8+ was reported when consuming a gluten-containing diet compared with a GFD.47 Further studies are needed. Third, changes of gut microbiota, as detected in CD,48 might also occur in patients with NCGS. Finally, a further aspect potentially linking NCGS with IBS is that HLA-DQ8 transgenic mice sensitised by gliadin displayed an increased secretion of acetylcholine from the myenteric plexus, enhancing muscle contractility and epithelial hypersecretion. Gluten withdrawal reversed both abnormalities.49

Evidence from double-blind placebo-controlled trials

Consistent evidence indicates the existence of an overlap between IBS and gluten-related disorders. In fact, about 5% of patients with IBS tests positive for CD and, conversely, CD may present with typical IBS-like symptoms.50 ,51 Moreover, IBS-like symptoms occur in the majority of patients with NCGS,40 while about one-third of patients with IBS may have NCWPS.37 Although wheat is now established to be linked to IBS, the component(s) that actually trigger(s) symptoms remain unknown. In this line, the only way to confirm the possible role of gluten or wheat as causative factors of NCGS/NCWPS is a DBPC strategy.27 This is an expensive and time-consuming procedure and, therefore, it is not yet of routine use being confined to research setting.37–39 ,43 ,52 ,53 So far, few DBPC trials have been performed. Their design and results are shown in table 1. The findings are discordant with a variation from approximately 30% of 920 patients with IBS in a routine clinical setting being sensitive to wheat protein, of whom the majority has NCWPS associated with multiple food hypersensitivity,37 to greater symptoms induction overall with gluten or wheat protein,38 ,39 ,53 to gluten-specific responses for current feelings of depression but not for abdominal symptoms.38 ,53

Reasons for apparent heterogeneity of results require dissection. First, subject selection might be a factor. For example, contamination of the cohort with CD can unduly skew results. This is why the exclusion of CD by combined histological and serological assessment while consuming adequate gluten is so important. In this respect, a critical point is to decide whether patients carrying HLA-DQ2/DQ8 and showing increased IEL density should be excluded from DBPC. Since about 40% of patients with NCGS is HLA-DQ2/DQ8+ and shows an increased IEL density, their exclusion would represent a pre-selection bias. In addition, 30–90% with positive responses to wheat protein/gluten has elevated IEL density in several studies.54 In the study in which 30% of patients with IBS showed sensitivity to wheat, there was a high incidence of eosinophilic infiltration of the mucosa and epithelium, features not described in the other reports, suggesting a different cohort being investigated.37 Second, nocebo responses can be a problem in re-challenge arms as evident in most, but not all, cross-over studies reported above. In the Australian study, 3 of 37 patients had gluten-specific induction of symptoms, but none of those three had such specificity of responses when a further DBPC challenge was instituted.43 The Italian study had 3 of 61 patients who demonstrated unequivocal specificity of gluten-induction of symptoms.39 It would be interesting to see if this is reproducible on a further DBPC challenge. Third, the active product that was used for the challenge differed from carbohydrate-depleted wheat protein43 ,53 to purified gluten39 to whole wheat flour,37 ,52 in its dose from low39 ,43 to high,37 ,38 ,39 ,52 and in its form of presentation from capsule37 ,39 to food.38 ,43 ,52 Finally, other design features including method of assessing the end-points and sample size differed.

Based on these findings, it is reasonable to speculate that gluten and/or other wheat proteins (such as ATIs)30 can generate intestinal and extraintestinal manifestations in a subgroup of patients with IBS. This contention is supported by mechanistic studies demonstrating epithelial injury and innate and possibly adaptive immune activation in response to wheat proteins. However, gluten and wheat are not the only dietary proteins involved in IBS. Proteins derived from milk, yeast and soy maybe involved in some,17 ,37 and IgE-mediated food allergy and nickel allergy have been reported in a significant proportion of patients with IBS and NCGS/NCWPS.37 ,40

The development of the FODMAP story

Over many years, there have been multiple observations that ingestion of certain short-chain carbohydrates—lactose, fructose and sorbitol, and fructo-oligosaccharides and galacto-oligosaccharides—was able to induce IBS-like symptoms, and that their restriction in the diet was associated with apparent improvement in symptoms in some patients with IBS (as reviewed in detail22). These carbohydrates have several key features in common. They are small molecules, containing only 1–10 sugars, and hence are possible osmotically active substances in the lumen of the intestine. They are slowly absorbed in the small intestine if monosaccharides are not absorbed at all if they contain more than one sugar due to lack of suitable hydrolases. Hence, they are present in the small intestinal lumen for a prolonged time and do increase the intestinal luminal water content. Their malabsorption leads to their exposure to intestinal bacteria, which rapidly ferment them to release short-chain fatty acids and gases (hydrogen, carbon dioxide and, in some people, methane). Their effects on symptoms and gas production are also additive.

Two hypotheses were proposed: (1) the luminal distension evoked by FODMAPs was related to symptom generation; (2) in patients with IBS and its associated visceral hypersensitivity, reducing the intake of all those short-chain carbohydrates would optimally improve the symptom burden. This was different to previous dietary strategies in that only one or two species of those carbohydrates—for example, lactose in lactose malabsorbers, fructose alone or in combination with sorbitol or fructans in fructose malabsorbers—were restricted.

Structure and implementation of the low FODMAP diet

A dietary approach was designed to reduce the intake of all FODMAP groups, by finding in each food group low-FODMAP alternatives. Other adjuncts to reducing FODMAP intake, such as the use of lactase in food or orally to reduce lactose content of relevant foods, and the use of co-ingestion of glucose with food containing an excess of free fructose, were also proposed in patients with IBS.22 Knowledge of the FODMAP content of foods was patchy and limited, and an ongoing programme of detailed food analyses has corrected aberrant assumptions and filled in many gaps.55–57 Such information has been made readily available by an application (the Monash University Low FODMAP Diet App) that is updated regularly.

The diet has been implemented by education via a dietitian trained in the principles of the diet. The dietitian would tailor the dietary advice to the eating patterns of the individual, ensure nutritional adequacy and provide written information and where to find accurate digital information. After 4–6 weeks, no or minimal response in an adherent patient should then lead to the abandonment of FODMAP restriction. If there has been a good response, consideration is then given to reducing the level of restriction by graded reintroduction of previously restricted foods with a focus on specific FODMAP groups. The patient is encouraged in the long term to restrict only to the level that is needed for symptomatic comfort.

Evidence to support the low FODMAP approach

It is timely to critically review whether the FODMAP concept and the use of a low FODMAP diet in patients with IBS are supported by evidence.