Approach to patients with refractory coeliac disease

Introduction

Coeliac disease (CD) is an autoimmune response triggered by dietary exposure to gluten in those individuals who are genetically predisposed¹. It results in both gastrointestinal and extra-gastrointestinal symptoms. Typically, patients have diarrhoea, bloating, weight loss, and anaemia. It is a lifelong condition with a prevalence of around 1% in Europe and the US. The gold standard for diagnosis requires a duodenal biopsy showing villous atrophy and increased intra-epithelial lymphocytes². The only accepted treatment for CD is to completely remove gluten from the diet; this includes cereal, wheat, rye, and barley. Most cases of CD report symptom improvement within a few weeks of a strict gluten-free diet (GFD). Those who do not respond to this diet are said to have non-responsive CD (NRCD). The difficulty of maintaining gluten elimination means that the majority of NRCD is due to continued ingestion. An alternative diagnosis should be sought in those members of the NRCD group whose continuing symptoms appear to be unrelated to CD. An even smaller group are those with refractory CD (RCD), in which a GFD is being adhered to but the symptoms of malabsorption remain. On biopsy, they will have persistent villous atrophy despite at least 12 months on a GFD. RCD can be classified as primary (no response to GFD) or, more commonly, secondary (an initial response to GFD, followed by relapse). RCD can also be classified by its clonality of T-cell receptor (TCR) into RCD types 1 and 2. It is important to make a distinction between RCD types, as type 2 is associated with complications such as ulcerative jejunitis and enteropathy-associated T-cell lymphoma (EATL), giving it a poor prognosis. Identifying the different types of RCD is a complicated process involving polymerase chain reaction (PCR) to determine TCR clonality, histology of the small intestine, and intra-epithelial lymphocyte (IEL) phenotype and morphology (Table 1)³.

RCD type 2 is rare and therefore there are only a few randomised controlled trials to give us treatment recommendations. In some centres, a combination of prednisolone with a thiopurine has been used and shown to give good results for type 1 and variable results for type 2; a clinical improvement was reported in up to 75% of patients with RCD type 2⁴. A review by Malamut et al. looked at treatment with methotrexate or anti-tumour necrosis factor-alpha in 14 patients with RCD type 1 and 43 patients with RCD type 2 and found that some had a histological response to this treatment⁵. There have been statistically significant improved survival rates following the use of cladribine (2-chlorodeoxyadenosine, or 2-CdA), whereas the use of alemtuzumab (anti-CD-52 monoclonal antibody) has shown only marginal success⁶. Previously, the use of azathioprine and prednisolone was associated with progression to EATL, but Nasr et al. showed that the use of these drugs in combination was not only successful but also safe, especially where the cases of RCD type 2 were identified early³. The results showed 53% histological recovery and 56% transformation from oligoclonal to polyclonal gamma receptor population, and none of the patients developed EATL.

Non-responsive coeliac disease

NRCD requires clinical and histological diagnosis. Following 1 year of a strict GFD, these patients will still be complaining of symptoms like abdominal pain, diarrhoea, and tiredness. Blood tests may show low levels of iron, B₁₂, and folate, and small bowel biopsy will reveal villous atrophy⁷. The main objective for managing these patients is to find the cause of the NRCD. Those who are diagnosed with NRCD can be classified into one of four groups on the basis of their history, current symptoms, and investigations: (1) patients who are not adhering to the GFD, (2) patients whose symptoms mimic untreated CD but who in fact have a second diagnosis other than CD, (3) patients with CD and complications, and (4) patients with RCD, which comprises 0.1% to 1% of patients with CD^3,8.

Over 90% of NRCD cases fall into the first group, consisting of those who continue to ingest gluten either by choice or unintentionally⁹. Nevertheless, when someone presents with NRCD, it is crucial to exclude other diseases which can mimic CD; these include pancreatic insufficiency, lactose intolerance, small bowel bacterial overgrowth, inflammatory bowel disease, hypo-gammaglobulinaemia, tropical sprue, collagenous colitis, and adult-onset autoimmune enteropathy. A detailed investigation is required to rule out these conditions; firstly, it is necessary to make sure the patient is strictly adhering to the GFD. If there are raised CD serology antibody titres, then it can be assumed that there is still ingestion of gluten. Investigations include an upper gastrointestinal endoscopy with biopsy of the small intestine for light microscopy, coeliac serology including IgA antibodies to tissue transglutaminase (tTG) and endomysium and PCR for TCR monoclonality. Immunoglobulin levels, including IgA and IgG titres, are tested by some centres, followed by testing for IgG antibodies to tTG if there is IgA deficiency. Further testing includes HLADQ2 and DQ8 status, colonoscopy, and testing for lactose and fructose intolerance, small bowel bacterial overgrowth, and pancreatic insufficiency. Following this, if CD is confirmed and other causes for the continuing symptoms have been ruled out, a strict GFD must be adhered to and RCD can be considered as the diagnosis.

Refractory coeliac disease

RCD represents a small subset of the NRCD group who continue to have symptoms of malabsorption and have villous atrophy on small bowel biopsy despite not having ingested gluten for at least 1 year¹⁰. This subset of individuals will already have been investigated for other causes, as detailed above, and usually nothing else is found. The differentiation into primary and secondary RCD can then be made according to the onset of symptoms. Those with primary RCD will never have had a response to a GFD, whereas those with secondary RCD will have been initially responsive to dietary elimination of gluten but are now experiencing a return of symptoms. There is no time frame for this; some patients develop secondary RCD decades later. Classifying RCD as type 1 or 2 depends on the phenotype of the IELs. Type 1 has a normal IEL phenotype, and type 2 has an abnormal clonal population with loss of CD8 and expression of intra-cytoplasmic CD3 by IELs. Making this distinction between types 1 and 2 is key in being able to manage RCD appropriately. It also aids with prognosis, as RCD type 2 has a 5-year mortality of 55% compared with just 7% in RCD type 1⁴. Ulcerative jejunitis and EATL are the main complications that can occur in RCD, and EATL is responsible for the majority of deaths in patients with RCD. There is a 3:1 female-to-male ratio with RCD type 2, but this is reversed in EATL to result in males being more affected¹¹. HLA DQ2 homozygosity is a risk factor for both RCD type 2 and EATL¹².

Diagnosis

It can be difficult to diagnose RCD. The process involves clinical assessment of the patient. Pathological, histological, laboratory, and radiological findings can all help, but essentially it is a diagnosis of exclusion in many cases. Nasr et al.³ describe the strategy used in a tertiary centre for diagnosing RCD type 2:

Management of refractory coeliac disease type 2

There have been a variety of attempts to manage RCD type 2. A good clinical response has been seen when using budesonide for RCD, but there is not a clear effect on prognosis¹³. Nasr et al.³ found good results when combining prednisolone with a thiopurine, including azathioprine, mercaptopurine, or thioguanine, and summarised the other trials undertaken in different centres (Table 2)^3,6,12–23. However, the positive results from thiopurines and prednisolone are not echoed by everyone. The Mulder group reported no improvement and/or progression to EATL when treating patients with the same regimen⁴. Nasr et al. argue that their good results are affected by early detection of RCD type 2, close monitoring of patients, adherence to the treatment, and a multidisciplinary approach which uses dieticians and a histopathologist who understands RCD and the availability of PCR³. They also acknowledge that the percentages of aberrant IEL and clonal TCR population may have an effect on response to treatment. Other treatments include the following:

Conclusions

RCD is a rare but serious condition. Although relatively high numbers of patients present with continuing symptoms of CD despite following a GFD, the majority of them can be attributed to the NRCD group 1, consisting of those who continue to ingest gluten. It is extremely important to be thorough in history taking, examination, and investigations of those presenting with NRCD in order to identify the small group with RCD. Once RCD has been diagnosed, the process of classification into type 1 or 2 can be undertaken by using small bowel biopsy, PCR, IEL phenotyping, capsule endoscopy, and imaging. Knowing whether a patient has RCD type 1 or 2 gives the opportunity to estimate the prognosis more accurately and guides treatment options.

Various treatment methods have been used in different centres, and the treatment outcome is variable. RCD type 2 is not common; however, early diagnosis and treatment may play an important role in the prognosis. Future areas for research may help identify other treatment options, particularly for difficult-to-treat cases or cases that have progressed.