What effect has gluten had on mankind?

Although mankind may have existed in some progressive form for 2·5 million years it is only in the last 10 000 years that we have been exposed to wheat. Wheat was originally cultivated in the Fertile Crescent (South Western Asia) with a farming expansion that lasted from approximately 9000 to 4000 BC. Thus, it could be considered that wheat and therefore gluten is a relatively novel introduction to man's diet⁽ Reference Accomando and Cataldo ^{1 )}. Gluten is a high molecular weight heterogeneous compound (which occurs not only in the endosperm of wheat but also in rye and barley) that can be fractionated to produce α, β and γ peptides.

Prior to 1939 (and the outbreak of World War II) the rationing system had already been devised. This led to an imperative to try to increase agricultural production. Thus, it was agreed in 1941 that there was a need to establish a Nutrition Society. A meeting of workers interested in nutritional problems was convened by Sir John Orr and was held at the Royal Institution⁽ Reference Copping ^{2 )}. The main objective of the new Society was to provide a common meeting place for workers in various fields of nutrition. The very roots of the Society were geared towards necessarily increasing the production of wheat⁽ Reference Copping ^{2 )}. This goal was achieved and by the end of the twentieth century, global wheat output had expanded 5-fold^{( 3 )}.

Coeliac disease (CD) or gluten sensitive enteropathy is a state of heightened immunological responsiveness to ingested gluten in genetically susceptible individuals. CD was uncommon with the estimated incidence in the UK in 1950 being reported as one in 8000⁽ Reference Davidson and Fountain ^{4 )}. However, we and others have reported in the last decade that CD affects 1% of all adults⁽ Reference Sanders, Patel and Stephenson ^{5 )}. Adult presentations are now more frequent than paediatric (9:1). Contemporary patients with adult CD rarely present with symptoms suggestive of malabsorption, more commonly now describing subtle gastrointestinal symptoms (e.g. irritable bowel type symptoms)⁽ Reference Sanders, Hopper and Azmy ^{6 ,} Reference Sanders, Carter and Hurlstone ^{7 )}. Any gastrointestinal presentation of CD is now termed the typical form. However, patients presenting with no gastrointestinal symptoms but alternatively extra-intestinal manifestations and/or recognised associated conditions (e.g. neurological symptoms or reduced bone mineral density)⁽ Reference Hadjivassiliou, Kandler and Chattopadhyay ^{8 –} Reference Hopper, Hadjivassiliou and Butt ^{16 )} are called the atypical form.

Delays in CD diagnosis are common, with symptoms in patients prior to the diagnosis ranging from 4·5 to 9·0 years⁽ Reference Sanders, Hurlstone and Stokes ^{17 )}. Furthermore, for every adult patient diagnosed approximately eight cases are still yet to be detected. These observations have led to protagonists advocating an active clinical policy for case finding and even debating whether the general population should be screened for CD. The mantra is ‘if you look for CD you will find it’⁽ Reference Richey, Howdle and Shaw ^{18 –} Reference Azmy, Hurlstone and Kong ^{20 )}.

Individuals with CD may initially be recognised by using either/both endomysial antibody and tissue transglutaminase (tTG) antibody. The positive predictive value of these serological tests is in excess of 90%⁽ Reference Hopper, Hadjivassiliou and Hurlstone ^{21 –} Reference Hopper, Cross and Hurlstone ^{23 )}. However, a duodenal biopsy with the demonstration of villous atrophy is still considered mandatory⁽ Reference Hopper, Cross and Sanders ^{24 –} Reference Hopper, Hurlstone and Leeds ^{28 )}. The recent interest in SNP may herald the advent of novel genetic testing⁽ Reference Van Heel, Lude and Hunt ^{29 ,} Reference Hunt, Zhernakova and Turner ^{30 )}.

The nutritional management of coeliac disease

Individuals with CD have a high prevalence of nutritional deficiencies, reduced bone mineral density and an increased mortality and risk of malignancy⁽ Reference Hopper, Leeds and Hurlstone ^{31 )}. The cornerstone of treatment for CD is lifelong adherence to a strict gluten-free diet (GFD). For the majority of patients, a GFD leads to clinical and histological remission, normalisation of standardised mortality rate, a reduction in long-term health complications (i.e. osteoporosis) and in some studies, an improvement in psychological well-being and quality of life (QOL)⁽ Reference Zarkadas, Cranney and Case ^{32 ,} Reference Mustalahti, Lohiniemi and Collin ^{33 )}.

QOL improves after 1 year on a GFD and may be sustained in the long term. Patients with CD on a GFD have a reduced QOL compared with healthy controls but this is still an improvement from their undiagnosed state. This benefit is particularly noticeable in patients who have presented with typical symptoms. Furthermore, we recently observed that patients with CD may report a worse QOL than patients with ulcerative colitis⁽ Reference Barratt, Leeds and Robinson ^{34 )}.

A GFD can be a major and initially overwhelming undertaking. Compliance to a GFD may be compromised by a lack of education, social restrictions⁽ Reference Karajeh, Hurlstone and Patel ^{35 )}, limited medical support (doctors and dietitians), the absence of symptoms at the time of diagnosis or in screen detected patients. If symptoms persist while on a GFD the most common reason is inadvertent non-adherence. However, there are a number of other conditions associated with CD that can cause persisting symptoms⁽ Reference Evans and Sanders ^{36 )}. We have reported associations with pancreatic exocrine insufficiency⁽ Reference Leeds, Hopper and Hurlstone ^{37 )} and co-existing inflammatory bowel disease⁽ Reference Leeds, Hœroldt and Sidhu ^{38 )}.

Exploration of the putative biological disadvantages of coeliac disease

Three novel publications suggest an alarming progression in the prevalence of CD with a rise of 2% (since the initial screening studies)⁽ Reference Lohi, Mustalahti and Kaukinen ^{39 –} Reference Catassi, Kryszak and Bhatti ^{41 )}. What possible mechanism could be involved? The suggestion that increased wheat consumption may play a role is highly plausible and is supported by internationally observed trends. Another alternative is that the rise in CD is in line with increasing prevalence and trends reported for other autoimmune diseases⁽ Reference Lohi, Mustalahti and Kaukinen ^{39 )}. Gluten consumption may also have a bearing on the likelihood of coeliac patients developing complications and gastrointestinal malignancy⁽ Reference Biagi and Corazza ^{42 )}.

In an obesogenic environment it has been suggested that undetected CD may confer a benefit to individuals. We have recently shown that the mean total cholesterol was 4·84 mm (sd 1·2) in newly diagnosed adults with CD (n 100). Males had 21% lower and females had 9% lower mean total cholesterol in comparison with the general population (difference in age-adjusted mean total cholesterol −1·09 (95% CI −0·97, −1·21) mm; −0·46 (95% CI −0·24, −0·68) mm, respectively). There was no change in mean total cholesterol following a GFD. However, there was a small but statistically significant increase of 0·12 mm (95% CI 0·05, 0·18) in the mean HDL-cholesterol. Thus, there appears to be little benefit conferred to patients with undetected CD⁽ Reference Lewis, Sanders and Logan ^{43 )}.

Our most recent work pertains to the role of detecting CD in adult patients with Type 1 diabetes. We showed that at diagnosis adult Type 1 diabetes patients with undetected CD had worse glycaemic control (8·2% v. 7·5%, P = 0·05), lower total cholesterol (4·1 v. 4·9, P = 0·014), lower HDL-cholesterol (1·1 v. 1·6, P = 0·017), a higher prevalence of retinopathy (58·3% v. 25%, P = 0·02), nephropathy (45% v. 5%, P = 0·009) and peripheral neuropathy (42·9% v. 15%, P = 0·11). However, there was no difference in QOL (all P > 0·1) after 1 year on a GFD, and only the lipid profile improved⁽ Reference Leeds, Hopper and Hadjivassiliou ^{44 )}. This suggests that the institution of a GFD in patients with an already complex diabetic diet does not adversely affect QOL⁽ Reference Leeds, Hopper and Hadjivassiliou ^{44 )}.

The spectrum of gluten-related disorders

Until recently, the terms gluten sensitivity and CD were used synonymously in the literature⁽ Reference Hadjivassiliou, Williamson and Woodroofe ^{45 )}. Gluten was associated only with CD and wheat allergy; therefore, patients with gluten-induced gastrointestinal symptoms who produced normal values of tTG, IgE and showed normal histology were advised to continue integrating gluten foods into their diet, as gluten was not regarded as the cause for their condition⁽ Reference Rostami and Villanacci ^{46 )}.

However, the majority of those seeking medical attention for gluten-induced gastrointestinal symptoms do not have CD or wheat allergy. Kaukinen and co-workers reported in a study with ninety-four adults affected by abdominal symptoms after cereal ingestion, a prevalence of 9% with CD, 8% with latent CD and 20% with cereal allergy. In total 63% of the study subjects could not be classified as CD or as allergic, but were yet affected by gluten foods and clinically benefited from a GFD⁽ Reference Kaukinen, Turjanmaa and Mäki ^{47 )}.

Furthermore, an increasing observation noted by healthcare professionals and the food industry is that the number of patients consuming a GFD seems greatly out of proportion to the projected number of CD patients. Although official data are lacking, the growing popularity in the gluten-free market was estimated at $2·5 billion in the US alone for the year 2010⁽ Reference Sapone, Bai and Ciacci ^{48 )}, with marketers estimating that up to 15–25% of the American consumers want gluten-free foods⁽ Reference Painter ^{49 )}.

Subsequently, this has led to the development of a consensus document on new nomenclature and classification of three gluten induced and heterogeneous conditions: CD, wheat allergy and non-coeliac gluten sensitivity (NCGS), a form of gluten intolerance that neither meets the diagnostic criteria for CD nor those for wheat allergy⁽ Reference Sapone, Bai and Ciacci ^{48 )}. The diversity of these gluten-induced conditions indicates that the immune system handles gliadin in different ways, with the mechanism underpinning NCGS currently poorly understood and of great research interest. In CD, there is increased small intestinal permeability allowing for aberrant passage of antigens and the activation of both an innate and adaptive immune response, ultimately manifesting as small intestinal damage and concurrence of tTG autoantibodies. In contrast, NCGS patients show normal intestinal permeability and have been proposed to fulfil the criteria to stimulate only an innate immune response, but seemingly fall short of triggering the adaptive immune response, presumably, as they lack the associated machinery to do so, such as epithelial permeability and the susceptible genetic markers human leucocyte antigen (HLA) DQ2 or DQ8 pattern. As a result the inflammatory small bowel response is halted and mucosal manifestations absent, in addition to negative tTG and endomysial antibodies⁽ Reference Sapone, Lammers and Casolaro ^{50 )}.

NCGS is an umbrella term and can incorporate a wide range of symptoms including abdominal discomfort, bloating, pain or diarrhoea; or it may present with a variety of extraintestinal symptoms; these may include headaches and migraines, lethargy and tiredness, attention deficit syndrome and hyperactivity, autism and schizophrenia, muscular disturbances as well as bone and joint pain (Table 1).

Table 1. Symptoms and associations in gluten sensitivity

In an article by Verdu et al., NCGS is defined by ‘one or more of a variety of immunological, morphological, or symptomatic manifestations that may also be shared by CD and irritable bowel syndrome (IBS)’ and thus in some quarters referred to as ‘no man's land’ between these two conditions⁽ Reference Verdu, Armstrong and Murray ^{51 )}.

There is a growing body of information that suggests patients may experience IBS symptoms related to gluten ingestion, yet in the absence of overt CD. Our group has previously reported that antigliadin antibodies were present in approximately 12% of the general population and approximately 17% in IBS (both in the presence of a normal small bowel biopsy)⁽ Reference Sanders, Patel and Stephenson ^{5 ,} Reference Sanders, Carter and Hurlstone ^{7 )}. Kaukinen et al. noted that the majority of patients with cereal-induced gastrointestinal symptoms complained of bloating (83%), diarrhoea (63%) and abdominal discomfort (34%), symptoms seemingly consistent with the ROME diagnostic criteria for IBS: of those in whom CD was excluded, antigliadin antibodies were present in 40%⁽ Reference Kaukinen, Turjanmaa and Mäki ^{47 )}. This further supports the hypothesis of GS-IBS (Fig. 1)⁽ Reference Verdu, Armstrong and Murray ^{51 ,} Reference Ball, Hadjivassiliou and Sanders ^{52 )}. What this may suggest is that in susceptible individuals the consumption of gluten results in an immune response that is manifested by the production of antigliadin antibodies.

Fig. 1. (colour online) A model for the relationship between coeliac disease (CD), irritable bowel syndrome (IBS) and gluten sensitivity (GS)⁽ Reference Ball, Hadjivassiliou and Sanders ^{52 )}.

Wahnschaffe et al. have suggested that serum coeliac-associated antibodies and the HLA DQ2 or DQ8 pattern may be predictive of a response to a GFD in diarrhoea-predominant IBS⁽ Reference Wahnschaffe, Schulzke and Zeitz ^{53 )}. They investigated forty-one diarrhoea-predominant IBS patients who agreed to participate in a 6-month course of a GFD. Positive IgG antigliadin antibodies/tTG and HLA DQ2 were all independent factors for a symptomatic response to a GFD.

Conversely, an intriguing study by an Australian group adds more to the debate⁽ Reference Biesiekierski, Newnham and Irving ^{54 )}. The investigators undertook a double-blind, randomised, placebo-controlled re-challenge trial in patients with IBS fulfilling the ROME III criteria (n 34). The patients were recruited through a newspaper advertisement and were already taking a self-imposed GFD due to symptomatic gastrointestinal benefit. CD was excluded on the basis of a negative HLA DQ2 or DQ8 pattern, or a normal duodenal biopsy (on gluten challenge) in the presence of a positive HLA. Patients were randomised according to a computer-generated list of random numbers held by an independent observer to either the gluten or the placebo treatment group. Over the 6-week study period, the severity scores of pain, satisfaction with stool consistency and tiredness were significantly higher for those consuming the gluten diet compared with the placebo group. Despite this, there was no evidence of intestinal inflammation or damage while being challenged with gluten (using faecal lactoferrin and intestinal permeability). Thus, these patients could not be viewed as having latent CD. This is the first study to describe an NCGS that may cause IBS⁽ Reference Biesiekierski, Newnham and Irving ^{54 )}.

In conclusion, we have demonstrated that with the progressive increase in gluten consumption this has led to an increased prevalence of CD with the associated complications. Furthermore, we have categorised the nutritional effects of a GFD in both newly diagnosed CD patients and Type 1 diabetes patients with previously undetected CD. Our recent work suggests that adult patients either with CD alone or CD in association with Type 1 diabetes derive benefit from a GFD. Finally, the increasing global use of a GFD and our recognition of a spectrum of conditions described as gluten-related disorders has opened up further thought-provoking avenues, whereby NCGS should not be seen as a ‘no man's land’ but a ‘fertile crescent’ for further research⁽ Reference Ball, Hadjivassiliou and Sanders ^{52 )}. Further work is required to understand the prevalence, clinicopathological features and serological biomarkers of NCGS, thereby potentially allowing the identification of which patients with IBS may derive benefit from a GFD.