In this article
Source: Benjamin Lebwohl, David S Sanders, Peter H R Green: The Lancet 2018, 391: 70-81
COELIAC DISEASE
Coeliac disease affects about 1% of the population. The frequency of diagnosis appears to be increasing, mainly because its incidence is genuinely rising. Coeliac disease develops in genetically susceptible individuals under the influence of unknown environmental factors. An immune reaction develops that is triggered by the digestion of gluten. The clinical presentation of the disease is highly variable: from severe malabsorption to mild symptoms or no symptoms at all. Diagnosis requires the demonstration of villous atrophy in the duodenum and the appearance of tissue transglutaminase antibodies in the circulation. In children, European guidelines allow biopsy to be avoided if characteristic symptoms and strict serological criteria are present. Although the disease can be treated effectively with a gluten-free diet, symptoms persist or return in a small proportion of patients. Because the diet is difficult to follow and because of refractory cases, more and more non-dietary treatments are being tested.
The disease primarily affects the small intestine, but its symptoms are wide-ranging: both intestinal and extra-intestinal symptoms can be observed.
Causative factors
Gluten
The main environmental factor responsible for the development of coeliac disease is gluten. Gluten is the prolamin storage protein of cereals (wheat, rye, barley). Gluten is the most important component in bread-making, giving bread its elasticity; however, it is rich in glutamine and proline, so gastric, pancreatic and brush-border peptidases can digest it only partially, and peptides as long as 33 amino acids may remain intact. These peptides enter the internal environment by the transcellular or paracellular route, where in susceptible individuals they trigger an adaptive immune reaction against the tissue transglutaminase (TTG) enzyme that deamidates gliadin peptides. TTG is the most characteristic autoantigen of the disease. Deamidation increases the immunogenicity of gliadin and helps the molecule bind to HLA-DQ2 or HLA-DQ8 molecules on antigen-presenting cells. The gliadin peptides are thus presented to CD4+ T cells. During the process, antibodies against TTG, gliadin and actin appear through an unknown mechanism. These antibodies may contribute to extra-intestinal symptoms such as dermatitis herpetiformis or gluten ataxia.
Alongside the adaptive immune response, activation of the innate immune response can also be observed in the intestinal epithelium, characteristically evidenced by an abnormal increase in intra-epithelial lymphocytes positive for NKT receptors (NKG2D and CD9/NKG2A). These receptors can recognise the stress-induced MICA and MICB gene products and the HLA-E protein that appears on the surface of epithelial cells. Up-regulation caused by IL-15 plays an important role in the appearance of these receptors. Both the adaptive and the innate immune response are needed for the full picture of coeliac disease to develop. The interaction between the two processes is not yet clear.
The role of non-gluten wheat proteins in the disease process is not clear. One group of these proteins, the amylase trypsin inhibitors, may contribute to damage to the intestinal epithelium by provoking the innate immune response, including in non-coeliac gluten sensitivity.
Genetic factors
The disease clusters in families. Almost 100% of patients carry specific variants of the HLA class II genes HLA-DQA1 and HLA-DQB1, which together encode the two (α and β) chains of the coeliac disease-associated heterodimeric proteins HLA-DQ2 and HLA-DQ8. These appear on the surface of antigen-presenting cells. More than 90% of patients are HLA-DQ2 positive, and almost all of the rest are HLA-DQ8 positive. Some individuals with coeliac disease do not carry the full complement of alleles of the HLA-DQ2 haplotype, so they are DQ2 negative but half-DQ2 positive. For this reason it is not enough to report the full haplotype; the presence of the allele components must also be tested.
Almost 100% of patients are HLA-DQ2, HLA-DQ8 or half-HLA-DQ2 positive. 40% of the European, American and South-East Asian population carry these alleles. Recent GWAS studies have identified 39 non-HLA region-associated genes that increase the risk of coeliac disease.
Environmental factors
Since the HLA genes mentioned are fairly common, yet the disease occurs in only 1% of the population, it is likely that environmental factors other than gluten also play a role in the development of the condition.
Breastfeeding and infant feeding practices
A change in infant feeding practices was suspected to lie behind the coeliac disease epidemic in Sweden between 1984 and 1996. No link with a change in coeliac disease risk could be shown either for breastfeeding or for the timing of gluten introduction among people with familial clustering or a characteristic HLA haplotype. An association with the development of the disease was observed only among those consuming larger amounts of gluten. Other studies, however, did not support even this. According to the ESPGHAN (European Society for Paediatric Gastroenterology Hepatology and Nutrition) guidance, gluten should be introduced into the diet between 4 and 12 months of age, and large amounts should be avoided in the first weeks after its introduction. A meta-analysis showed that introducing it later than 6 months of age increases the risk of coeliac disease.
Other risk factors
Incidence may be affected by the season of birth and by whether an elective caesarean section was performed – although the effect of the latter is not clear. Gastrointestinal infection (rotavirus in children, Campylobacter jejuni infection in adults) is a risk factor. The total number of infections in the first 18 months of life (more than 10 compared with fewer than 4) and respiratory infection in the first 18 months of life increase the risk. Similarly, taking antibiotics or proton pump inhibitors increases the risk, as does iron treatment of the mother during pregnancy. Infection with the otherwise non-pathogenic reovirus can trigger the disease. Helicobacter pylori colonisation may reduce the risk.
The role of the microbiome
Cross-sectional studies show that the microbiome changes in coeliac disease and does not fully normalise even on a gluten-free diet. The faecal concentration of Bifidobacterium bifidum was significantly higher in untreated coeliac patients than in healthy people, and Gram-negative, potentially pro-inflammatory bacteria were more common in the duodenum of children with coeliac disease than in the control group. The gut bacterial composition of vaginally born, breastfed, at-risk HLA-DQ2 infants was specific, suggesting that HLA type plays a role in shaping the composition of the microbiome.
Epidemiology
The prevalence of coeliac disease in the population is 1%, but differences that cannot be explained by genetic and environmental factors are observed worldwide. Within Europe, prevalence is lower in Germany than in other European countries. Prevalence is highest in Sweden and Finland. In the United States, prevalence is lower among the African American population than among white people, and in Brazil it is lower among indigenous Brazilians and people of African descent. Prevalence is rising in India and is highest in northern India. In this region the frequency of HLA haplotypes is similar to that in the rest of India, but wheat consumption is higher. According to a large study carried out in the USA in 2016 (500,000 duodenal biopsies), the frequency of villous atrophy among people originating from the northern Indian state of Punjab was 3.08%, compared with 1.80% in the rest of the US population.
While the prevalence of the disease is increasing, the frequency of diagnosis is rising much more slowly. It is estimated that 80% of people with coeliac disease were undiagnosed by 2009, and that this proportion decreased by less than 50% by 2013–14.
Clinical presentation
The frequency of presentation has shifted from the classic childhood form (malabsorption) towards non-classic forms in both children and adults. Classic symptoms – diarrhoea, weight loss, faltering growth – are relatively rare nowadays. Non-classic symptoms are much more common: iron deficiency, bloating, constipation, chronic fatigue, headache, abdominal pain and osteoporosis. Greater attention must be paid to non-classic symptoms and to timely diagnosis. According to a UK survey, the average delay in diagnosis was 13 years.
Terminology describing patients with coeliac disease according to Ludvigsson and colleagues (2013):
| Potential | Positive serological test, normal biopsy result |
| Asymptomatic | Symptoms are absent even when specific symptoms are asked about |
| Symptomatic | Gastrointestinal or extra-intestinal symptoms |
| Classic | Diarrhoea, signs of malabsorption or both |
| Non-classic | No malabsorption, but other symptoms present (e.g. anaemia, osteoporosis) |
| Refractory | Symptoms and villous atrophy persist despite a gluten-free diet. |
When testing should be carried out to confirm or rule out coeliac disease according to NICE (National Institute for Health and Care Excellence) guidelines
Testing for coeliac disease is recommended
- Persistent abdominal or gastrointestinal symptoms that cannot be explained otherwise
- Faltering growth
- Unexpected weight loss
- Severe or persistent mouth ulcers
- Unexpected iron, vitamin B12 or folic acid deficiency
- Type 1 diabetes mellitus
- Autoimmune thyroid disease
- Irritable bowel syndrome (IBS)
- First-degree relative of a patient with coeliac disease
Testing for coeliac disease should be considered
- Metabolic bone disease (osteoporosis or osteomalacia)
- Unexplained neurological symptoms (mainly peripheral neuropathy or ataxia)
- Unexplained raised temperature, recurrent miscarriage
- Persistently raised liver enzyme levels of unknown origin
- Dental enamel defects
- Down syndrome
- Turner syndrome
Diagnosis
Diagnosis requires serological tests and a biopsy from the duodenum. For screening people in risk groups, testing for TTG IgA antibodies is recommended first. Its sensitivity and negative predictive value are high, and it is less costly than testing for endomysial antibodies (EMA). If TTG IgA is only mildly raised, the highly specific EMA test should also be performed. (The latter is more expensive and labour-intensive, requires special experience on the part of the assessor, and its evaluation is subjective. In some patients, IgA against deamidated gliadin peptide (DGP IgA) is positive while TTG IgA is negative.
In IgA-deficient patients the IgA tests (TTG, EMA, DGP) will be negative; in such cases IgG antibodies should be tested.
Coeliac disease can only be ruled out if the patient was on a gluten-containing diet before testing. Historically, at least 10 g a day (about 4 slices of bread) had to be eaten for 6 weeks before testing. This rule has since been relaxed (10 g of gluten a day for 18 days, or more than 3 g of gluten for at least 2 weeks). After eating 3 g of gluten a day for two weeks, villous atrophy develops in 68% of patients, while serological markers rise more slowly (becoming positive in half of patients after 14 days and in 89.5% after 28 days). If the pre-test probability of coeliac disease is low, a negative serological test performed on day 28 after the start of the 2-week diet is sufficient to rule out coeliac disease. If the pre-test probability of the disease is high, it is definitely advisable to perform a biopsy at the end of the 2nd week after starting the diet.
In the duodenal sample, the combined presence of an increased number of intraepithelial lymphocytes, crypt hyperplasia and villous atrophy (Marsh 3), together with positive serology, confirms the diagnosis of the disease. This histological picture can also occur in other conditions (e.g. giardiasis or HIV infection), and conversely, the duodenal sample may appear negative despite positive serology. Performing a biopsy is essential for a definite diagnosis before the patient starts a gluten-free diet. This baseline sampling is also important so that a baseline biopsy result is available for patients who do not respond to the gluten-free diet. Improvement on a gluten-free diet does not prove a diagnosis of coeliac disease, since improvement can also be observed in non-coeliac gluten sensitivity (NCGS). A negative serological result on its own does not rule out coeliac disease with 100% certainty.
In paediatrics, the 2012 ESPGHAN (European Society for Paediatric Gastroenterology, Hepatology and Nutrition) guideline stated that in symptomatic paediatric cases a duodenal biopsy is not necessary if the TTG concentration exceeds 10 times the upper limit of normal, EMA is positive in a sample obtained from a repeat blood draw, and the patient has HLA type DQ2 or DQ8. This procedure rules out errors arising from (1) sample labelling mistakes and (2) transient autoimmunity. For adult patients there are not yet enough data for a diagnostic algorithm that omits biopsy to be considered reliable.
A reliable histological diagnosis requires an appropriate sampling strategy, since atrophic and non-atrophic areas may alternate in the duodenum. According to the current position, one sample should be taken from the bulb and a further 4 from the second part of the duodenum. Material taken from the bulb must be handled separately (because Brunner’s glands and heterotopic gastric mucosa occur in the bulb, the pathologist needs to know this sampling site). The technique that allows histological orientation is the so-called single-bite biopsy (this technique is needed to measure the crypt/villus ratio properly).
HLA genotyping is not part of routine diagnostic procedures, but it can be valuable when the diagnosis is uncertain, especially in people who are already on a gluten-free diet and do not wish to undergo a gluten challenge, or who refuse a small-intestinal biopsy. The absence of HLA-DQ2 or DQ8 almost always rules out coeliac disease; its negative predictive value is more than 99%. Precise determination of the allele components is necessary, since coeliac disease can develop even with half-DQ2, although this is rare. For those who refuse gastroscopy, video capsule endoscopy is a further option, which makes changes in the mucosa visible.
In individuals whose serological result is negative but whose histology has shown villous atrophy, besides the possibility of another disease, they may (in 2–15% of cases) have seronegative coeliac disease. This can occur in people with IgA deficiency, other immunodeficiency states or those receiving immunosuppressive medicines. The condition can occur at the onset of the disease or in its late stage (in the presence of enteropathy-associated T-cell lymphoma), or if the patient is already on a gluten-free diet. People with seronegative coeliac disease were more often older (49 years compared with 36 years), had characteristic symptoms (e.g. diarrhoea), showed more severe villous atrophy and also had other autoimmune diseases. Despite seronegativity, anti-TTG2 antibodies could often be detected in the villous biopsy samples. According to some views, seronegative coeliac disease is a more severe form of the disease.
Diagnosing seronegative coeliac disease requires great care: alongside HLA-DQ2 or DQ8 positivity, the exclusion of other conditions causing villous atrophy and the positive effect of a gluten-free diet confirm the condition. According to a survey of 200 seronegative cases, 31% of patients turned out to have coeliac disease, 27% had another diagnosis, and 18% were found to be idiopathic – 72% of the latter recovered spontaneously, without any intervention. Most seronegative cases were observed among non-white people, in whom the cause was mainly infection. Other causes of seronegative villous atrophy were certain medicines (mycophenolate mofetil and olmesartan medoxomil), common variable immunodeficiency, Crohn’s disease, tropical sprue, giardiasis, autoimmune enteropathy and other rarer diseases.
Controversies in coeliac disease testing
A beneficial effect of screening on the outcome of the disease has not yet been proven.
Treatment
The mainstay of treatment is a gluten-free diet. Improvement can be observed within days to weeks, serological markers normalise and villous atrophy disappears. Following a gluten-free diet, however, is not easy (price, quality). Adherence can easily slip when travelling or eating out. The presence of traces of gluten in various foods is uncertain. At diagnosis, a dietitian’s advice should always be sought, not only to learn about hidden sources of gluten but also to set up a healthy diet with adequate fibre and nutritional value. In many cases, trace element and vitamin deficiencies (iron, folic acid, vitamin B12, vitamin D) need to be corrected at diagnosis. Pneumococcal vaccination is also recommended, as patients are more prone to community-acquired pneumonia. For people over 55, bone densitometry is recommended 1 year after starting the gluten-free diet because of the higher frequency of osteoporosis and bone fractures. After the diagnosis has been made, regular dietary and symptom monitoring is needed.
Non-dietary treatments
Numerous non-dietary treatment options are being tried, either as alternative or as complementary solutions. Approaches based on the following principles are being tried:
- inactivation of toxic peptides inside the gut
- preventing gliadin from passing through the intestinal wall
- induction of immune tolerance
- inactivation of immune processes in the lamina propria
- use of a vaccine containing gluten-specific CD4-positive T cells.
Two medicines are currently at the experimental stage:
- Larazotide acetate: a peptide that, taken by mouth, modifies the tight junction (TJ) and prevents gliadin peptides from passing through – (used alongside a gluten-free diet) it has so far proved better than placebo.
- Latiglutenase: an enzyme that prevents gluten-induced damage – it has so far not proved better than placebo.
Both medicines require further study.
Non-responsive and refractory coeliac disease
About 20% of patients complain that their symptoms persist or return despite a gluten-free diet. Non-responsive coeliac disease can have several causes:
- The initial diagnosis was not correct:
- non-coeliac gluten sensitivity (NCGS)
- seronegative villous atrophy not caused by coeliac disease
- Unintentional gluten intake
- Other coexisting conditions
- Irritable bowel syndrome (IBS)
- Small intestinal bacterial overgrowth (SIBO)
- Other food intolerance (e.g. lactose intolerance, fructose malabsorption)
- Exocrine pancreatic insufficiency
- Microscopic colitis
- Refractory coeliac disease
In such cases the first step is to confirm that the initial diagnosis was accurate; HLA typing and reassessment after a gluten challenge may be needed.
If the diagnosis of coeliac disease is clear and the complaints persist, accidental gluten intake should be suspected and a dietitian experienced in the field should be consulted. Other coexisting diseases can also cause symptoms to persist (IBS, microscopic colitis, lactose intolerance, fructose malabsorption, pancreatic insufficiency, SIBO).
During follow-up, the patient may continue to have villous atrophy, with or without symptoms. This is seen mainly in people over 50, but villous atrophy was also detected in 19% of children who had a repeat biopsy after 1 year. In adults with persistent villous atrophy, fractures associated with osteoporosis and malignant lymphoproliferative diseases occurred more often. This raises the question of whether an intestinal biopsy should be performed during follow-up. Persistent villous atrophy calls for closer monitoring and an even stricter diet (eliminating all processed foods).
Refractory coeliac disease can be diagnosed if symptoms persist or return and villous atrophy does not resolve even though the patient has demonstrably been on a strict gluten-free diet. Such conditions carry an increased risk of mortality, and enteropathy-associated T-cell lymphoma (EATL) develops more often. In such cases a repeat duodenal biopsy is needed, in which the abnormal intraepithelial T-cell population is further examined using immunohistochemistry, flow cytometry and PCR (analysis of T-cell receptor – TCR – γ gene rearrangement). In type 2 refractory coeliac disease, CD8, CD3 and TCRγ expression is absent from the surface of the aberrant intraepithelial T cells, while intracytoplasmic CD3 can be detected and oligoclonal TCR gene rearrangement can be demonstrated. Patients with refractory coeliac disease in whom no aberrant T-cell population can be detected (type 1 refractory coeliac disease) have a better prognosis. In type 2 coeliac disease there is a high risk that the process will progress to EAT lymphoma.
Risk of malignancy and mortality
Besides the EATL already mentioned, other types of non-Hodgkin lymphoma are also more likely to develop in people with coeliac disease than in people without the disease. In addition, a higher frequency of other malignant tumours of the small intestine, large intestine, liver, pancreas and oesophagus has also been suggested, but apart from the first year after the diagnosis of coeliac disease, only malignancies of the small intestine and liver remained significant. In a subsequent study, the frequency of colorectal cancer also proved to be moderately increased more than 5 years after diagnosis. At the same time, the risk of lung and breast cancer was found to be lower in the coeliac population, perhaps because the proportion of smokers in this population is significantly lower. According to a more recent meta-analysis (2012), there is no association between coeliac disease and malignant diseases overall.
Numerous studies support the increased mortality of people with untreated coeliac disease. This mortality decreases once adequate treatment is introduced. According to a Swedish study, the increased mortality is due to a higher frequency of cardiovascular disease, lung disease and cancer. What can certainly be concluded from the conflicting studies is that the mortality risk of people with untreated coeliac disease is moderately higher than that of the population without the disease.