The gluten‐free diet: an historical perspective and its use by people without coeliac disease

The long term benefits and risks of gluten avoidance for people without coeliac disease are unknown

Interest in food and its connection with disease dates back to ancient medicine.1 Despite scientific and technological advances, diet modifications continue to have a role in medicine, and certain foods are directly involved in some conditions or symptoms.2,3

Coeliac disease is one such condition. It was first described in ad 2, but it was not until 1888 that a dietary treatment was recommended, and 1950 that gluten was specifically implicated.4 Coeliac disease is the only autoimmune disorder for which the environmental trigger is known, and a gluten‐free diet is the only treatment. Patients classically present during childhood with diarrhoea, weight loss, and complications of malabsorption, such as vitamin and mineral deficiencies. However, health care providers are now also cognisant that patients may present with constipation or extra‐intestinal symptoms (eg, fatigue), that they can develop coeliac disease well into adulthood, and that they can be overweight or obese at diagnosis. The increased awareness of coeliac disease among health care providers has been matched by that of the general public, leading to greater commercial availability of (more palatable) gluten‐free foods. In fact, retail sales of gluten‐free products doubled between 2011 and 2016 and are expected to continue growing worldwide.5

Interest in the gluten‐free diet for people without coeliac disease has developed since it was first discussed in the medical literature in 1980: the symptoms of eight women with indications consistent with coeliac disease, but with otherwise unremarkable findings on evaluation, resolved during a gluten‐free diet.6 Increased awareness of coeliac disease and the availability of gluten‐free products prompted interest in its use for treating gastrointestinal and extra‐intestinal symptoms, as well as for general health. In a nationwide survey in the United States during 2009–2014, the proportion of people without diagnoses of coeliac disease following gluten‐free diets increased from 0.5% to 1.7%, while the coeliac disease diagnosis rate was stable.7

Patients follow gluten‐free diets for a variety of reasons: to lessen the gastrointestinal symptoms of nausea, bloating, abdominal pain, or diarrhoea; to relieve extra‐intestinal symptoms of fatigue, headaches, difficulty in concentrating, and arthralgias; for weight loss or a perceived general health benefit; and to support household members with coeliac disease. Yet the long term benefits and risks of a gluten avoidance for people without coeliac disease are unknown.

Studying the effects of diet on health is challenging for several reasons. As foods are chemically complex, it can be difficult to pinpoint which components trigger symptoms in some people. The emergence of the low fermentable oligo‐, di‐, mono‐saccharides and polyols (FODMAP) diet best highlights this challenge, as participants with self‐reported non‐coeliac gluten sensitivity (NCGS) may continue to have symptoms despite strict adherence to the gluten‐free diet.8 NCGS is particularly difficult to study because of the heterogeneity of its clinical presentation, the lack of a diagnostic test, and its poorly understood pathophysiology. Further, symptomatic response to treatment is variable, and there is considerable overlap between people with NCGS and those with functional gastrointestinal disorders. Thoughtful studies are needed to better guide health care providers and patients regarding the utility of gluten‐free diets for people without coeliac disease, paying particular attention to appropriate patient and symptom selection and the duration of the gluten‐free diet trial.

Distinguishing NCGS from coeliac disease is becoming clinically challenging as more people start gluten‐free diets without first seeking medical advice. Objective markers and methods for distinguishing NCGS from coeliac disease in patients on gluten‐free diets will be essential for both clinical practice and designing investigations. Recent studies have found elevated plasma levels of interleukin‐2 after a single gluten challenge, and that double blind, sham‐controlled gluten challenges can distinguish between people with coeliac disease or NCGS, even after prolonged adherence to a gluten‐free diet.8,9,10

As reported in this issue of the MJA, Potter and his colleagues11 examined the incidence of self‐reported NCGS, its resolution, and the association of each with functional gastrointestinal diagnoses in a general population sample of adults responding to two successive Digestive Health and Wellbeing postal surveys. About the same proportion of people reported NCGS in 2015 and 2018; 41% of those who reported NCGS in 2015 did not report it in the 2018 follow‐up survey. Of the people who reported NCGS in 2018, 13.9% followed strict gluten‐free diets (a further 57.2% partially avoided gluten).

These findings have several possible explanations: NCGS may spontaneously resolve without intervention; gluten is not the cause of symptoms in people with self‐reported NCGS; the side effects or burden of the gluten‐free diet did not outweigh its benefits. Recognising the adverse effects of a gluten‐free diet is an important consideration: data in this area are limited, but vitamin deficiencies, particularly thiamine deficiency, and elevated arsenic levels are possible in people on gluten‐free diets.12,13

In summary, a gluten‐free diet has long been recognised as the only treatment for people with coeliac disease. Over the past four decades, increasing numbers of people without coeliac disease but with gastrointestinal or extra‐intestinal symptoms have adopted gluten‐free diets. Robust studies with careful phenotyping of participants are needed to better define the natural history of NCGS, and its relationship with gluten or wheat needs to be investigated. Avoiding gluten will substantially affect other dietary components, such as FODMAPs, that may be alternative explanations for reported symptoms. Such a radical change in diet will also influence the gastrointestinal microbiome, with unknown consequences. Studies of the relationships of foods with human syndromes are often inadequate for determining causality, and their findings should not be the basis for recommending major dietary changes in humans.