Gluten-related disorders: certainties, questions and doubts

Abstract

In the last decade, the ingestion of gluten, a heterogeneous complex of proteins present in wheat, rice, barley and probably in oats, has been associated with clinical disorders, such as celiac disease, wheat allergy and recently to non-celiac gluten sensitivity or wheat intolerance syndrome. Gluten-related disorders, which are becoming epidemiologically relevant with an estimated global prevalence of about 5%, require the exclusion of gluten from the diet. For the past 5 years, an important shift in the availability of gluten-free products, together with increased consumption in the general population, has been recorded and is estimated to be about 12–25%. Many people follow a self-prescribed gluten-free diet, despite the fact that the majority have not first been previously excluded, or confirmed, as having gluten disorders. They rely on claims that a gluten-free diet improves general health. In this review, we provide an overview of the clinical disorders related to gluten or wheat ingestion, pointing out the current certainties, open questions, possible answers and several doubts in the management of these conditions.

KEY MESSAGE

• Incidence of gluten-related disorders is increased in the last decade and self-diagnosis is frequent with inappropriate starting of a gluten-free diet.

• Gluten and wheat are considered as the most important triggers to coeliac disease, wheat allergy and non-celiac gluten sensitivity.

• Pediatricians, allergologist and gastroenterologist are involved in the management of these conditions and appropriate diagnostic protocols are required.

Keywords:

• Gluten
• non-celiac gluten sensitivity
• wheat allergy
• wheat intolerance syndrome

Introduction

There is a unified school of thought that the term “gluten-related disorders” is an umbrella term used to describe all conditions related to the ingestion of gluten or wheat-containing food [1]. Celiac disease (CD), wheat allergy (WA) and non-celiac gluten sensitivity (NCGS) or wheat intolerance syndrome (WIS), whose prevalence has increased in the last decade, are all classified among the gluten-related disorders. CD is an immune-mediated gluten-dependent systemic disorder characterized by a well-established serologic and genetic profile, along with small intestinal damage [2]. WA occurs when the body’s immune system reacts to one or more of the proteins found in wheat. It occurs in children as well as adults, but is usually outgrown in infancy [3] with IgE- and non-IgE-mediated mechanisms. Recently, a growing number of papers have reported a link between types of food consumption and functional gastrointestinal (GI) symptoms, such as abdominal pain, diarrhoea and dyspepsia [4]. Wheat is considered as a major input in determining these symptoms, and all its components may be involved (gluten proteins, α-amylase/trypsin inhibitors, fermentable oligosaccharides, disaccharides, monosaccharides and polyols), in the absence of diagnostic biomarkers [5]. This paper gives an updated overview on “wheat or gluten” related clinical conditions with the aim of discussing certainties and controversies that can help physicians to improve diagnostic accuracy.

Wheat structure and gluten components

Wheat is among the most important food plants used for human alimentation, along with corn and rice, and is cultivated in all continents of the world. Cultivated wheat comprises diploid, tetraploid and hexaploid species. Diploid einkorn (Triticum monococcum) and tetraploid emmer (Triticum turgidum) were among the earliest cultivated forms of wheat derived from wild grasses, and probably appeared about 10,000 years ago, in the south-eastern area of Turkey [6,7]. About a thousand years later, in the Middle East, a hexaploid wheat, which originated from hybridization between cultivated emmer and a related wild grass, appeared for the first time [7]. Currently, the most widely cultivated species are tetraploid (Triticum durum) and hexaploid (Triticum aestivum) wheat variants, which, respectively, originated from durum wheat (also called pasta wheat), mainly cultivated in Mediterranean countries (Italy, France, Spain, Greece), and from soft wheat (also called bread wheat), which is the most common variety of wheat cultivated in the rest of the world (about 95% of wheat grown worldwide), because of its better adaptability to different meteorological conditions. The wheat grain is a single-seeded fruit called a caryopsis. The caryopsis of wheat is structured in three parts: the fibre part called bran, composed of three layers (from outside to inside: pericarp, seed coat and aleurone layer), which represents the envelope of the caryopsis (14% of the structure), the embryo or germ (3% of the caryopsis, rich in fats, proteins and vitamins) and the endosperm, rich in starch and proteins, constituting 80% of the caryopsis (Figure 1) [8]. About 85% of the caryopsis consists of carbohydrates, of which about 80% is represented by starch, 12% by cell wall polysaccharides and 8% by low molecular mass; mono-, di- and oligosaccharides, including fructans (fructo-oligosaccharides) [9,7]. They are classified among fermentable oligosaccharides, disaccharides, monosaccharides and polyols (FODMAPs) and considered as one of the possible causes of intestinal and extra-intestinal symptoms in NCGS [10]. Although the quantity and quality of starch and lipids in wheat caryopses contribute to the derived flour characteristics, proteins are the most important components of caryopsis and determine the features of flours and derived doughs. Albumin, globulin, gliadin and glutenin are the principal wheat proteins. Prolamins (gliadin and glutenin) represent about 80–90% of wheat caryopsis proteins, and are localized in the starchy endosperm cells; these storage proteins are the most important components of gluten and influence its suitability for bread-making of the flour obtained from wheat. Gliadin and glutenin, in contact with water alongside the mechanical action of kneading, form covalent (disulphide bonds) and non-covalent bonds (hydrogen bonds, ionic bonds, Van der Waals forces) and assemble a protein complex called gluten, which determines the extensibility, elasticity and strength of doughs derived from wheat flour. Specifically, gliadins determine the extensibility, whereas, glutenins influence the elasticity and strength of the gluten. Different types of prolamins are also contained in other cereals, such as barley (hordeins) and rye (secalins) [11–14]. Other caryopsis proteins include proteins associated with cell structures and enzymes. These proteins, which belong to inhibitors of trypsin and the α-amylase family, constitute two-thirds of caryopsis albumins and are involved in IgE-mediated mechanism of WA and could be regarded as a possible cause of NCGS symptoms [15].

Figure 1. Structure of the caryopsis of wheat.

Celiac disease

CD is an immune-mediated reaction to a triggering environmental factor (gluten) in genetically susceptible individuals. The prevalence of CD is approximately 1% in the North American and European populations, and may be higher in Northern European countries, approximately 1.5% [16]. Similar to many chronic inflammatory diseases and autoimmune diseases, CD is substantially increasing in prevalence [17]. In many developing countries, the frequency of CD is likely to increase in the near future, given the diffuse tendency to adopt Western, gluten-rich dietary patterns [18,19]. The pathogenesis of CD is characterized by an inappropriate T-cell-mediated immune response, causing self-perpetuating inflammatory damage of the small bowel and other organs, with variable combination of gluten-dependent clinical manifestations, in genetically predisposed subjects [4]. HLA-DQ2 (DQA1*05/DQB1*02) is associated with most cases of coeliac disease, whereas, HLA-DQ8 (DQA1*0301/DQB1*0302) is present in just a minority of patients [20]. Gluten, in genetically predisposed people, is the most important environmental factor in the pathogenesis of CD. When gluten is introduced in the digestive tract, gliadins and glutenins are partially hydrolyzed by proteases localized in the GI tract [21]. Gluten-derived peptides, arguably after the increased permeability of the tight junction [22], reach the lamina propria where gluten-peptides are presented by HLA-DQ2- and HLA-DQ8-positive antigen-presenting cells, and thus, drive the immune response. The glutamine and proline residues of the gluten-peptides induce them as a substrate for tissue transglutaminase (tTG), which potentiate presentation by HLA-DQ2 and HLA-DQ8 by deaminating or transamidation gluten peptides [23]. In addition, cross-linking between gliadin and tTG determines the formation of new epitopes; they both trigger the primary immune response and determine the development of auto-antibodies against tTG [24]. These immunogenic epitopes are presented by HLA-DQ2 or HLA-DQ8-positive antigen-presenting cells to CD4+ T-cells. HLA expression on antigen-presenting cells is supported by local production of IFNγ of CD4+ T-cells [25]. CD4+ T-cells, producing interleukins, such as IL-15, IL-2 and IL-21, mediate an inflammatory process leading to mucosal injury in the small bowel consisting of a redistribution of the intraepithelial lymphocytes in the epithelium at the tip of the villi, including epithelial cell killing and remodelling [26] (Figure 2). There is a modest gender bias favouring females [27]. Generally, first-degree family members of patients with CD have an increased risk of the disease (ranging from 2% to 20%), depending on gender and HLA haplotype [28]. Patients with other autoimmune diseases or selective immunoglobulin A deficiency, and those with Down syndrome, Turner syndrome and Williams syndrome, should undergo blood tests for the diagnosis of CD as well as those having an increased risk of developing it [29]. CD has a variable clinical presentation and symptoms are frequently unrecognized, thus, it is possible that for each child diagnosed with CD, about 3–7 cases remain undiagnosed [30]. CD diagnosis is a challenge for clinicians because of its heterogeneous clinical picture, including GI-related symptoms (diarrhoea, weight loss, abdominal distention and constipation) and extra-intestinal or atypical manifestations (growth retardation, iron deficiency anaemia, chronic fatigue, weight loss, aphthous stomatitis, delayed puberty, amenorrhea, reduced bone mineral density or alterations in liver function tests) [2]. Due to a greater awareness of symptom diversity, extra-intestinal manifestations are now more recognized in children than before [29]. In addition, CD may be asymptomatic: these patients are often diagnosed through testing of populations enrolled during screening programs or in case-finding strategies for detecting CD in patients with disorders that are associated with a high risk of CD [31] as in the case of 43% of the children identified by family screening [28]. The typical form of CD, with classical intestinal symptoms of malabsorption, may be diagnosed at any age of life, presenting positive serological biomarkers and villous atrophy as a histological feature. The atypical form is characterized by extra-intestinal symptoms and positive celiac serology with limited abnormalities of the small intestinal mucosa and no malabsorption signs. The silent form of CD is defined as a condition characterized by detection of positive CD-specific antibodies (anti-endomysium antibody and/or anti-tTG), HLA predisposing gene, small-bowel biopsy findings compatible with CD, but the absence of typical clinical GI symptoms; in this case, as in typical and atypical forms, the therapeutic option is represented by a gluten-free long-life diet. The latent form is characterized by the presence of predisposing genes, HLA-DQ2 and/or HLA-DQ8 without enteropathy, in the presence and/or absence of symptoms and CD-specific antibodies [29]. The potential form of CD includes the presence of celiac serology positivity and compatible HLA, in the absence of intestinal damage; the patient may or may not have symptoms and signs, and may or may not develop a gluten-dependent enteropathy later [29]. In these last forms, latent and potential, periodical screening for the disease is recommended. Tosco et al. conducted a prospective, 3-year cohort study to determine the natural history of potential CD in children. They enrolled 106 patients with potential CD based on serology analysis, HLA-DQ2 and/or DQ8 haplotype and normal duodenal architecture; a 6-month clinical and serological screening was performed. Duodenal biopsy analysis was repeated after 2 years in patients with persistent positive serology and/or symptoms. Eighty-nine of the 106 patients entered the follow-up study, with normal daily consumption of gluten, while the others started a GFD as they had symptoms. After 3 years, approximately 33% of patients developed villous atrophy. Most children with potential CD remained healthy [32]. Kurppa et al. showed, in a randomized controlled study, that all symptomatic patients with mild enteropathy may benefit from a GFD and that all the patients left on a gluten-containing diet developed atrophy after 1 year of observation [33]. Specific CD autoantibodies are detected in the serum against tTG, endomysium (EMA) and deamidated gliadin peptides (DGP) [29]. The serological tests used in children >2 years are tTG and EMA antibodies of class A with high sensitivities (98% and 90%, respectively) and specificities (97% and 98%, respectively), especially in patients with severe histological small bowel alterations [34]. Total immunoglobulin measurement is also important as CD is associated with selective IgA deficiency [35]. The combination of tTG and DGP antibodies of class G has shown itself to be particularly useful in patients with IgA deficiency and children <3 years of age [36]. HLA-typing is not a routine diagnostic test of CD given that 40% of the general population present the relevant haplotypes. HLA-typing is useful in selecting individuals at risk of the disease, such as first-degree family of affected patients because of its very high negative predictive value [37]. The characteristic histological pattern of the small bowel mucosa in patients with CD is partial to total villous atrophy with crypt hyperplasia and intraepithelial lymphocytic infiltration, rated according to the Marsh–Oberhuber classification, ranging from type 0 (normal) to 4. A Marsh–Oberhuber classification type 3 (a, b or c), or type 2 if accompanied by specific symptoms of CD supports the diagnosis. The severity of clinical symptoms does not correlate with the severity of histological alterations [30]. Four of these five criteria should be satisfied in the diagnosis of CD: (a) presence of symptoms associated with CD, (b) presence of CD-associated autoantibodies (tTG, EMA, DPG), positivity for HLA-DQ2 or -DQ8 alleles, (c) duodenal biopsy demonstrating blunting or absence of villi with >25 lymphocytes/100 enterocytes and (d) improvement of symptoms after a gluten-free diet [38]. Recently, European Society for Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) guidelines have proposed that it may be possible to avoid intestinal biopsy in children who meet the following criteria: symptoms consistent with CD, serum tTG antibodies of class A levels >10 times the upper limit of normal (confirmed with positive anti-EMA in a different blood sample) and positive HLA-DQ2 and/or HLA-DQ8. In all other cases, small bowel biopsies are still mandatory for diagnosis [29]. CD screening is commonly characterized by serum anti-tTG and/or EMA antibody detection levels: serum presence of these was demonstrated in culture supernatants of oral, duodenal and colonic biopsy samples [39]. Literature also reports that the so-called celiac intestinal antibody pattern (anti-gliadin, EMA and anti-tTG antibodies) can be highlighted in gut lavage fluid samples from CD patients [40]. Anti-tTG detection in saliva from CD patients has opened the possibility to use this biological fluid for diagnostic purposes [41], or monitoring antibody trends after starting a GFD [42]. At present, these investigations have not been validated for diagnostic purposes [43]. Picarelli et al. have reported the presence of EMA and anti-tTG antibodies in the faecal supernatants of CD patients, suggesting that intestinal mucosa is the site of CD-specific antibody production [44]. Recent studies suggest the possibility of in vitro diagnosis with the use of duodenal mucosa organ culture, especially when clinical data, serology and histology are not sufficient for the diagnosis of CD [45,46]. Di Tola et al. used immunochromatographic assay to demonstrate that anti-tTGs are detectable in duodenal culture media of most CD patients secondary to the intensity of indicative lines of the anti-tTG concentration. Immunochromatographic assay, despite lower sensitivity and diagnostic accuracy, is proposed in their report as a new diagnostic tool of in vitro CD diagnosis [47]. Carroccio et al. identified the organ culture system as a valid procedure to highlight gluten-specific immunological activation, showing that EMA and anti-tTG antibodies are detectable in the supernatants of cultured duodenal biopsies from CD patients [48,49]. Picarelli et al. evaluated the occurrence of CD-specific immunological activation in the colonic mucosa using the organ culture system. They demonstrated the presence of EMA and anti-tTG antibodies in the supernatants of cultured sigmoid biopsy samples from CD patients, confirming that the immune response to gluten can involve several GI tracts [39]. The urgency to correctly diagnose and devise effective interventions is driven by recent evidence of incidence of certain types of cancers among CD patients. This includes non-Hodgkin lymphoma, enteropathy-associated T-cell lymphoma, small intestinal adenocarcinoma, and esophageal and oropharyngeal squamous carcinoma [50,51]. With regard to GI neoplasms, it was noted that CD patients have a higher risk of developing small bowel adenocarcinoma, with an estimated odds ratio (OR) ranging from 4.29 to 59.97 [52]. In a meta-analysis, Han et al. evaluated the association between CD and risk of malignancies; the results indicated that CD increased the risks of all malignancies as well as GI malignancy, including esophageal cancer and small intestinal carcinoma, specifically. Compared with the general population, the risk of all malignancies was slightly increased in CD patients (pooled OR  = 1.25, 95% CI: 1.09–1.44) [53]. The only therapy for CD is lifelong adherence to a GFD that improves clinical complaints, restoring mucosal alterations and reducing the risk of complications [30]. Siegel et al. have proposed the use of enzymes as oral supplements to enhance gluten degradation, thereby reducing gluten exposure [54]. Kapoerchan et al., in another study, explored the possibility of blockers in order to prevent gluten peptide binding to HLA-DQ2 [55]. Klock et al. studied the inhibition of tTG, which prevents gluten modification and CD4+ T cells response [56]. Larazotide acetate, an oral peptide capable of improving small bowel barrier function, was tested in order to prevent the crossing of gluten peptides into the lamina propria [57]. Clinical response with symptom improvement is considered as the diagnostic tool in CD. CD antibodies are sensitive and specific for diagnosis but during follow-up they do not reflect dietary adherence or persistence of enteropathy [58]. Recent studies have proposed new possible serological markers of CD or possible supporting markers of damage such as serum intestinal-fatty acid binding protein (I-FABP) [59] or anti-actin autoantibodies (AAA) [60]. In the study by Adriaanse et al., the accuracy of I-FABP, a marker for intestinal epithelial damage, was investigated to predict intestinal damage during gluten challenge. In this study, serum I-FABP levels increased within 14 days in the CD patients after the introduction of gluten in the diet, correlating, both at diagnosis and after two weeks of gluten exposure, with intraepithelial lymphocyte count [61,62]. The increase of I-FABP is more rapid than CD antibody titres. These data suggest that serum I-FABP can be an early, non-invasive marker of gluten-induced enteropathy in CD, becoming an important tool for clinical and research settings [62].

Figure 2. Antigen presentation, inflammatory process and antibody production in celiac disease. Gluten peptides, after crossing the intestinal epithelium, reach lamina propria where they are deamidated by the enzyme tissue-transglutaminase (tTG) and are presented by HLA-DQ2 or HLA-DQ8-positive antigen-presenting cells (APC) to CD4 + T cells. CD4 + T cells activate both an inflammatory process leading to small bowel mucosal injury (TH1 reaction), and stimulate B cells to produce antibodies against tTG, deaminated gluten peptides, gluten peptides-tTG complex (TH2 reaction).

AAA have also been proposed as a marker of CD activity. AAA–IgA are directed toward the intracellular cytoskeleton actin filaments [60,63]. Schirru et al., in a retrospective study, hypothesized that the findings of positivity for AAA–IgA in children and adolescents with anti-TG2 antibody levels between 4 and 10 times the upper limit of normal, could in future be used to reduce the number of duodenal biopsies [64]. AAA–IgA level is considered as a reliable marker of intestinal mucosal damage in the CD patients and a simple ELISA test can be carried out. Achour et al. reported an AAA–IgA prevalence of 41.3%, in a celiac population, with a specificity of 71.4% showing diminished AAA–IgA levels in children strictly following a gluten-free diet (GFD) [65]. The need for non-invasive approaches to monitor CD and evaluate compliance with a GFD is certainly warranted. Moreno et al. proposed the detection of gluten immunogenic peptides in human samples, such as faeces and urine, as new non-invasive biomarkers to detect gluten intake and verify GFD compliance in the CD patients [66].

In summary, the certainties are that CD is one of the best-understood HLA-linked disorders with a defined trigger (gluten proteins from wheat and related cereals). The question of why HLA-DQ2 or HLA-DQ8 are expressed in 30–35% of the populations where CD is prevalent and only 2–5% of gene carriers develop CD is still open. The possible answer is that other genetic, as well as environmental, factors are involved as contributors. The remaining doubts are on the role and development of new intraluminal therapies focused on reducing gluten immunogenicity.

Wheat allergy

WA is a food allergy disorder to one or more of the proteins found in wheat, which can be either IgE-mediated, mixed IgE or non-IgE (cell) mediated or even only non-IgE (cell) mediated [67]. It is one of the most common food allergies in children, especially in early childhood [68]. Epidemiological studies report a prevalence of WA in children of about 0.4% [69] and familiar history of atopy is a risk factor in developing an IgE-mediated food allergy in children, as in other allergic diseases [70]. Recently, Nwaru et al. conducted a systemic review and meta-analysis on European prevalence of food allergies for all age groups reported to have WA, after cow’s milk and egg allergy, as the third most frequent self-reported food allergy, with a prevalence rate of 3.6%, but with only 0.1% prevalence of verified WA by food challenge [71]. IgE-mediated WA causes skin, respiratory and GI tract involvement and in most severe cases, anaphylaxis [72]. This may also occur as a wheat-dependent exercise-induced anaphylaxis (WDEIA) due to the combination of wheat ingestion before physical activity even if other co-factors have been described, such as temperature (warm or cold) and drugs, especially nonsteroidal anti-inflammatory drugs [73]. Baker’s asthma and rhinitis, caused by inhalation of raw wheat flour, depend also on an IgE-mediated reaction, as demonstrated by Houba et al. [74]. Another major occupational form of IgE-mediated WA is contact urticaria, and also forms of non-IgE-mediated occupational hand contact dermatitis may occur [75]. IgE-mediated allergic contact urticaria caused by hydrolyzed wheat in cosmetics has also been reported [76]. The most involved allergens are alfa-amylase inhibitors, but gliadins have been included as important allergens in the pathogenesis of WA, and different classes seem to be involved in the IgE response [77,78], such as ω-5 gliadins and high-molecular-weight glutenin subunits, in WDEIA [79]. The symptoms of WA develop within minutes to hours after gluten ingestion, typical for an IgE-mediated allergy, and include itching and swelling in the mouth, nose, eyes and throat; skin rash or swelling; wheezing in the respiratory tract; GI symptoms such as cramps, bloating and diarrhoea; and life-threatening anaphylaxis [80]. With regard to other food allergies, the gold standard for diagnosis of WA is a double-blind, placebo-controlled food challenge, but it is associated with a risk of anaphylaxis. In clinical practice, symptom resolution, in response to dietary elimination of wheat, usually confirms a suspected diagnosis of WA [81]. Skin prick tests (SPT) and wheat-specific IgE testing or, to a lesser extent, component-specific-IgE are the first-level studies for diagnosis of IgE-mediated WA, but they are not an alternative to food challenges. Soares-Weiser et al. reported sensitivities of 83% and 73% for specific-IgE and SPT, respectively; the presence of wheat-specific IgE seems to have a lower specificity than SPT (43% and 73%, respectively), as wheat-specific IgEs are detectable in most atopic children [82]. The second possible allergic reaction to wheat proteins is a delayed antibody-independent T-cell response which is a non-IgE-mediated reaction to different wheat protein fractions [68]. Atopic eczema/dermatitis is a form of non-IgE-mediated WA, which can be investigated with an oral double-blind food challenge [83]. Mixed IgE- and non-IgE-mediated food allergic reactions have been described as eosinophilic, esophagitis and gastroenteritis; a chronic inflammatory disease of the digestive system [84]. Kalgawalla et al. reported that wheat was identified as responsible for esophagus inflammation in 26% of patients (both adults and children) screened with SPT [85]. In a paediatric population, the removal of foods identified on SPTs or atopy patch tests (ATPs), such as milk, egg, wheat and soy, led to histologic resolution in 77% of patients [86]. In delayed non-IgE-mediated responses to wheat, detection of IgE sensitization to wheat is not helpful for diagnosis: APT can be used because of its high level of specificity, but it has a weak sensitivity [82]. Subjects with WDEIA must avoid exercise for about 4 h after wheat ingestion [80]. In comparison with other food allergies, the prognosis for WA is relatively good: the majority of children outgrow their allergy at 12 years of age [3]. In those who manifest an anaphylactic reaction to wheat, dietary elimination should probably be permanent, although loss of sensitization has been documented [81]. WA and CD do not usually occur in the same subject, but the case of a 4-year-old girl has been reported with IgE-mediated allergy to cereals, including wheat, and latent CD [87,88].

The certainties are that wheat can determine allergy to IgE-mediated, non-IgE-mediated or a mixture of both mechanisms. These clinical conditions are different from CD, which is a T-cell-mediated (type 4 hypersensitivity) immune response to gluten. An unresolved point is lack of a well-accepted set of criteria for the diagnosis of WA. Knowing that the detection of food-specific IgE does not indicate a diagnosis of WA, the possible answer to this point is that the diagnosis of WA always requires an oral challenge test. The remaining doubts are the already described cases of the diagnosis of WA in CD, but the relationship between these two conditions is unclear.

Non-celiac gluten sensitivity or wheat intolerance syndrome

NCGS is the term used to describe a clinical condition in which intestinal and extra-intestinal symptoms are related to gluten ingestion, improving or disappearing within hours or a few days after gluten withdrawal, and in which both CD and WA are completely ruled out. The symptoms usually relapse following gluten reintroduction [4]. The prevalence of this condition is unknown, especially in children. Indirect evidence suggests that NCGS affects up to 5–10% of Western population [89] being slightly more common than CD. NCGS has been mostly described in adults, particularly in females within the 30–50 year age group; however, paediatric case series have also been reported [90,91]. Some studies have reported normal intestinal permeability, or minimal histological alterations compatible with Marsh–Oberhuber 0–1 classification [92]. Sapone et al. reported increased permeability in a subgroup of HLA-DQ2/DQ8-positive adult patients [93]. With a higher prevalence than the general population, that is 30–40%, half of the patients with NCGS are HLA-DQ2 or -DQ8 positive [4]. Gene expression analyses showed an increased expression of Toll-like receptor 2 (TLR2) and reduced expression of the regulatory-T-cell marker FOXP3 in patients with NCGS when compared to patients with CD, suggesting a pathogenic role for innate immunity [30]. Contrary to CD, however, most studies show that adaptive immunity markers are not increased in patients with NCGS [93]. Intestinal irritable bowel syndrome (IBS)-like symptoms are nonspecific, including abdominal pain, bloating and diarrhoea. Extra-intestinal symptoms, less frequent in children, are fatigue, headache, joint and/or muscle pain, weight loss, anaemia, dermatitis and behavioural disturbances [91]. Carroccio et al. have recently demonstrated a similar incidence of autoimmune disorders between CD and NCGS patients [94]. The gluten-related peptides result in an innate immune response and an additional adaptive immune response with increased expression of IL-6, IL-21, IL-17 and IFN-γ [54]. In patients with suspected diagnosis of NCGS, gliadins do not induce mucosal inflammation or basophil activation as demonstrated in CD [95]. There are no biological markers specific to NCGS. Volta et al. evaluated the presence of anti-gliadin antibodies of class G and A occurring in 56% and 8% of the patients, respectively, in comparison to 80% and 75%, respectively, in the CD population [96]. In particular, anti-gliadin antibodies of class G diminish after a GFD. However, anti-gliadin antibodies are also frequently present in the general population. Rodrigo et al. demonstrated a similar incidence of anti-gliadin antibodies of class G and A among patients with NCGS and gluten ataxia compared to the CD patients; thus, some authors have suggested that gluten ataxia may be a clinical expression of NCGS [97]. Recently, Hadjivassiliou et al. reported a similar incidence of gluten ataxia, peripheral neuropathy and encephalopathy, and equal improvement in neurological symptoms on a GFD among the CD and NCGS patients. Therefore, they suggested the use of every available serological testing for patients suspected of having gluten-related disorders, also considering the important pathogenetic role of isoenzymes tTG2 and tTG6, both in CD and NCGS [98]. Lau et al. suggested an increased immune reactivity to gluten in children with autism, demonstrating significantly higher levels of IgG antibody to gliadin compared with the controls, particularly among children with autism and GI symptoms. Authors proposed a pathogenetic mechanism, different from CD, which may involve gluten in the pathogenesis of autism [99]. However, two reviews reported no strong evidence on the efficacy of gluten in an exclusion diet in the improvement of autistic spectrum disorder symptoms [100,101]. In addition to gluten, another cause of persistent symptoms is multiple food intolerance/hypersensitivities [102]. Other grain proteins may be responsible for symptoms in NCGS. Junker et al. have hypothesized that α-amylase trypsin inhibitors (ATI), activators of innate immune responses contained in wheat and related cereals, can give a clinical picture of NCGS [15]. Biesiekierski et al. demonstrated the possibility that other molecules, different from gluten and other wheat proteins, such as FODMAPs can be the cause of symptoms typical of NCGS [10]. The FODMAP list includes fructans, galactans, fructose and polyols that are contained in several foodstuffs, including wheat, vegetables and milk derivatives. Hence, it should be stressed that FODMAPs cannot be entirely and exclusively responsible for the symptoms reported by NCGS subjects, since these patients experience a resolution of symptoms while on a GFD, despite continued ingestion of FODMAPs from other sources, such as legumes [90]. A systematic elimination and re-introduction of suspected items, guided by dietitians, is necessary to confirm the diagnosis of FODMAP intolerance [103].

A double-blind, placebo-controlled (DBPC) challenge has been suggested to confirm NCGS diagnosis [104]. It is possible to use an open food challenge, but there is a higher probability of false positive results because of important placebo effects [105].

Biesiekierski et al. [92] proposed a randomized DBPC challenge to establish the possible relationship between gluten ingestion and intestinal and extra-intestinal symptoms, and to support the presence of NCGS, in patients with IBS diagnosis in whom CD was excluded and symptoms improved on a GFD. Patients on a GFD (started at least 6 weeks before the challenge) were randomized into two groups; the first group consumed a “food-vehicle” containing 16 g of wheat gluten in total every day, the second one consumed a food-vehicle without gluten. The vehicles had the same texture and taste and were free from FODMAPs. At baseline, the end of each study week and three weeks after completion of the challenge, patients completed a symptom questionnaire and a visual analogue scale for overall and specific symptoms. Authors demonstrated a significantly greater worsening of symptoms between patients who consumed a gluten-containing diet, both weekly and at the end of the challenge. Carroccio et al. [106], in a retrospective study, reported a DBPC challenge performed in a group of patients with IBS-like symptoms. Patients after a gluten reintroduction, for at least two weeks, later underwent a standard elimination diet with exclusion of wheat, cow’s milk, eggs, tomato and chocolate for the other 4 weeks. Each food was reintroduced one at a time; in the case of wheat reintroduction, patients consumed a capsule containing wheat or xylose for two consecutive weeks and, after one week of washout, they received the other capsule for the other 2 weeks. During all the phases of the study, patients completed a visual analogue scale for overall and specific symptoms. Authors demonstrated a significant and progressive worsening of symptoms only in patients on a wheat-containing diet. However, a limitation of this study was that they did not use gluten but wheat for the challenge, therefore they cannot exclude that other components of wheat (such as FODMAPs) could be responsible for symptoms.

In another randomized DBPC trial, Biesiekierski et al. [10], in a cohort of patients with suspected NCGS, in whom CD was excluded, presented three diets differing in gluten content, and, in order to control other potential factors responsible for intestinal symptoms, all diets had a reduced and controlled content of FODMAPs, dairy products and food chemicals. The first part of the challenge consisted in 2 weeks of GFD low in FODMAPs, successively, 1 week of high-gluten (16 g/d whole-wheat gluten), or low-gluten (2 g/d whole-wheat gluten and 14 g/d whey protein isolate) or placebo (16 g/d whey protein isolate) diet, followed by a washout period of at least 2 weeks and until symptoms induced during the previous dietary challenge resolved, before crossing over to the next diet. The second part of the challenge (rechallenge trial) was a 3-day cross-over challenge, conducted and programmed after the results of the first part of the challenge were analysed. It consisted of receiving one of the three dietary treatments (containing 16 g/d whole-wheat gluten or 16 g/d whey protein isolate or no additional protein), in which “food-vehicles” were similar in texture and taste. Authors demonstrated that symptoms, evaluated by a visual analogue scale, significantly improved during low FODMAPs diet and significantly worsened when gluten or whey protein was reintroduced; they did not prove a dose-dependent gluten influence on symptoms.

Recently, during the International Experts Meeting on gluten-related disorders, experts proposed that a diagnostic protocol should be performed for the confirmation of NCGS [107]. The primary outcome should be a clinical response to a GFD and the secondary outcome monitoring the effects of gluten reintroduction. The challenge consists of two steps, during which clinical evaluation is performed weekly using a self-administered Gastrointestinal Symptom Rating Scale (GSRS), to report symptoms, and a Numerical Rating Scale (NRS), to rate the intensity of relevant symptoms. In the first step, in patients on a GFD for at least 6 weeks, after CD and WA exclusion, the response to a GFD is verified and only patient “responders” (defined as patients who presented >30% reduction of one to three main symptoms or at least one symptom with no worsening of others, for at least 50% of the observation time) will continue to the second step. The second step is a DBPC challenge with crossover (in clinical practice a single blinded procedure could be sufficient) consisting of a one-week challenge (8 g/day of gluten or placebo) followed by a one-week washout of strict GFD and by the crossover to the second one-week challenge. Gluten-vehicles should contain 8 g of gluten and an exact quantity of ATI, and it should be FODMAPs free and undistinguishable from placebo in look, texture and taste.

Picarelli et al. [108] hypothesized, on the basis of the possible similarity between small intestinal and oral mucosa, that gluten may determine an allergic contact mucositis-like condition in NCGS patients on a gluten containing diet, therefore they carried out a DBPC gluten oral mucosa patch test (GOMPT) in patients with suspected NCGS, comparing results with treated CD patients, untreated CD patients and healthy controls. Authors demonstrated a significantly higher incidence of positive results of GOMPT for local and systemic symptoms (intestinal and extra-intestinal symptoms) in NCGS patients compared to the other three groups. Therefore, authors suggested GOMPT as a valuable tool in the NCGS diagnostic work-up. In contrast to CD or WA, NCGS remains a controversial entity with more questions than answers concerning its pathogenesis, diagnosis and treatment. Randomized controlled clinical trials in children with GI symptoms are necessary to evaluate the pathogenesis and treatment of NCGS, avoiding unnecessary gluten diet exclusion that may have negative health effects (Figure 3). The certainties are that wheat is considered as one of the foods known to evoke IBS symptoms. The question on which components of wheat are mostly responsible for clinical symptoms remains an unresolved point. The possible answers to this question are that gluten proteins can alter gut physiology, resulting in visceral hypersensitivity and abnormal motility. The remaining doubts concern the search for methods to detect these sensitivities, accurately. We propose a diagnostic flow-chart for patients with suspected gluten-related disorders according to a position statement and a clinical report published by the Italian Association of Hospital Gastroenterologists and Endoscopists (AIGO) [109], the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN) [61] and the Experts’ agreement and recommendations of the Salerno Experts’ Criteria [107] (Figure 3).

Figure 3. Diagnostic flow-chart in suspected gluten-related disorders. *According to Salerno experts’ criteria. **Positive challenge is defined as a variation of at least 30% of one to three main symptoms between the gluten and the placebo challenge. Gluten-free diet (GFD), human leukocyte antigen (HLA), celiac disease (CD), wheat allergy (WA), fermentable oligosaccharides, disaccharides, monosaccharides and polyols (FODMAPs), non-celiac gluten sensitivity (NCGS).

Conclusions

The GFD approach is indicated in CD and WA. Paediatricians, allergologists and gastroenterologists are involved in the diagnosis and follow-up of these conditions. Recently, a high number of studies has linked type of food consumed with GI symptoms. Gluten exposure is considered as the most important etiological factor with the description of new entities such as NCGS and WS. Self-diagnosis of these conditions is frequent in the general population, with inappropriate and prearranged elimination of gluten from the diet. In conclusion, the certainties are that gluten and non-gluten components of wheat are responsible for triggering GI and systemic symptoms outside the diagnosis of WA and CD (Figure 4). The persistent unresolved question is what diagnostic biomarkers should be considered for NCGS diagnosis. The physicians must be able to provide their patients with more certainties and less concerns or controversies.

Figure 4. Characteristics of celiac disease, wheat allergy and non-celiac gluten sensitivity. Human leukocyte antigen (HLA), anti-tissue transglutaminase antibodies (tTG), anti-endomysial antibodies (EMA), anti-gliadin antibodies (AGA), fermentable oligosaccharides, disaccharides, monosaccharides and polyols (FODMAPs), α-amylase/trypsin inhibitors (ATI).

Contributors’ statement

Claudio Romano, Simona Valenti, Domenico Corica and Luisa Ricciardi conceptualized this study, carried out data acquisition and interpretation, drafted the article and approved the final manuscript as submitted. They agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Disclosure statement

None of the authors have any conflicts of interest to disclose that are relevant to this article.

None of the authors have any financial disclosures to make that are relevant to this article.