A small subset of patients with celiac disease become refractory to a gluten-free diet, with persistent or recurrent symptoms of malabsorption and intestinal villous atrophy. This condition, defined as refractory celiac disease (RCD), is diagnosed after other small bowel diseases with villous atrophy are excluded. RCD is subdivided into 2 subgroups: type I RCD and type II RCD (RCDII). This latter condition is considered a low-grade intraepithelial lymphoma and has a poor prognosis. This article reviews the clinical and pathologic features of RCD and recent pathogenic findings in RCDII, offering a model to study how inflammation can drive T-cell lymphomagenesis.
- •
Novel concepts of refractory celiac disease have recently emerged and refer to 2 distinct entities sustained by 2 different pathogenic mechanisms.
- •
Type I refractory celiac disease (RCDI) is indistinguishable from uncomplicated active celiac disease except in its autonomy toward gluten exposure.
- •
RCDII resembles a low-grade lymphoma characterized by clonal expansion of small aberrant intraepithelial lymphocytes (IELs).
- •
Diagnosis is based on specialized small bowel investigations (enteroscopy, videocapsule endoscopy) and techniques of IEL analyses (immunohistochemistry, molecular biology, flow cytometry).
- •
Survival of patients with RCDI has been described as being inferior to that of those with celiac disease, and is associated with passage of RCDI to RCDII and onset of overt lymphoma.
- •
Prognosis of RCDII is poor because of severe malnutrition and very high risk of overt lymphoma.
- •
Recent advances in understanding of the pathogenesis of both forms of refractory celiac disease (RCD) offer new targeted strategies to cure RCD and prevent overt lymphoma.
Background
Celiac disease is an enteropathy related to autoimmune diseases induced by gluten in genetically predisposed individuals with HLA-DQ2 and HLA-DQ8 genotypes. The rate of diagnosis has been rising dramatically, although, with a prevalence of 1% of Caucasians in Europe and the Americas, most cases remain undetected. Its clinical presentation is extremely variable, and diagnosis relies on the detection of specific serum antibodies and the demonstration of intestinal villous atrophy. Treatment relies on strict adherence to a lifelong gluten-free diet (GFD), which generally prevents bone, autoimmune, and malignant complications. An initial nonresponse to a GFD is mainly from accidental or deliberate gluten contamination of the diet. Once gluten contamination and other causes of symptoms and villous atrophy have been ruled out, a small subgroup of patients with celiac disease remain who may have a primarily or secondary resistance to a GFD because of an authentic RCD. This article describes the diagnosis and treatment of the different types of RCD to provide guidance for the care of this rare but serious condition.
Diagnosis of RCD
Definition
RCD can be defined as persistent or recurrent severe enteropathy and symptomatic malabsorption in patients with previously confirmed celiac disease who have been on a GFD for at least 12 months (see exception mentioned later) and whose enteropathy or symptoms are not otherwise explained by other conditions or gluten conatamination. A very small subgroup of patients with celiac disease develop primary or secondary resistance to a GFD. Incomplete exclusion of dietary gluten must first be eliminated, because as many as 50% of patients are less than optimally compliant. Persistent or recurrent symptoms of malabsorption with intestinal villous atrophy for at least 12 months of a strict GFD characterize RCD. Diagnosis of this condition is made after other small bowel diseases are excluded, such as autoimmune enteropathy, tropical sprue, or common variable immunodeficiency, and the other conditions that are highly likely to explain the symptoms ( Fig. 1 ). RCD has been subdivided into 2 subgroups: type I (RCDI), defined by persisting villous atrophy despite a strict GFD associated with an increased number of IELs bearing a normal phenotype with surface CD3 and CD8 expression, and type II (RCDII) characterized by clonal expansion of abnormal IELs lacking surface markers CD3, CD8, and T-cell receptors, and preserved expression of intracellular CD3.
Epidemiology
Frequency of RCDI and RCDII remain unknown. In the Derby cohort, West and Holmes report that 0.7% of 713 patients with celiac disease had RCDII. In this latest study, diagnosis of patients with RCDII was based only on aspects of ulcerative jejunitis, which could have either caused an underestimation of RCDII from lack of application of molecular techniques or overestimation from confounding with RCDI. Incidence of T-cell lymphoma associated with celiac disease has been estimated at 0.22 to 1.9 per 100,000 inhabitants per year, although some of these cases may proceed directly from celiac disease, often undetected, to lymphoma without passing through a clinically apparent syndrome of RCD.
Predisposing Factors
Whether patients with RCD have a particular genetic background differentiating them from those with uncomplicated CD is still debated. The small numbers of patients with RCD have had limited genetic investigations. However, severity of celiac disease was correlated with the number of HLA-DQ2 copies: homozygosity for HLA-DQ2 was observed in 25.5% of patients with RCDI, 44.1% of patients with RCDII, and 53.3% of patients with enteropathy-associated T-cell lymphoma (EATL), versus 20.7% of patients with uncomplicated celiac disease and 2.1% of controls. Other genes involved in lymphocyte signaling [genes: SH2B3 (12q34), PTPN2 (18q11), RGS1 (1q31)] are associated with celiac disease and could be involved in the risk of developing overt lymphoma. Ongoing genomewide association studies suggest that the known celiac susceptibility variants may be not found in RCDII.
Another important environmental factor is probably the exposure to gluten. Risk of lymphomatous complications was reported to be 4 times higher in patients not observing a GFD than in those who were compliant. The amount of gluten consumption could be responsible for the differences in terms of severity of celiac disease. A recent study shows that the more severe outcome of celiac disease in southern compared with northern Europe is related to a higher gluten intake.
Diagnosis
Diagnosis of RCD relies on persisting malabsorption and villous atrophy after 1 year of following a strict GFD typically confirmed by expert dietician review. An exception may be necessary for patients who present with severe malabsorption syndrome with ulcerative jejunitis who have strictly adhered to a GFD and are still experiencing an inexorable decline. These cases must be differentiated from the very severe presentation that can occasionally occur in regular celiac disease. Initial endoscopic assessment includes standard or extended upper gastrointestinal endoscopy with biopsy. Sufficient biopsies must be taken to allow for a full assessment of the degree of mucosal injury and for special studies to determine the type of RCD or alternative diagnoses. Careful visual assessment for focal abnormalities that could indicate a prevalent malignancy is essential. Most will be situated in the proximal small intestine but may be beyond the reach of the standard or even an extended endoscopy. Double-balloon enteroscopy is a superior tool for identifying suspicious lesions for EATL or for a better assessment of ulcers, particularly for evidence of ulcerative jejunitis found in roughly 70% of patients with RCDII. Capsule endoscopy is useful for showing the extent of lesions. Capsule endoscopy has superior sensitivity for predicting persistent villous atrophy compared with standard video endoscopy. Furthermore, capsule endoscopy allows the visualization of ulcers all along the intestinal tract, which may suggest RCDII before diagnostic confirmation through direct biopsy sampling. Moreover, the authors diagnosed 3 cases of overt lymphoma when capsule endoscopy showed very suspicious intestinal strictures and jejunal ulcers. Double-balloon enteroscopy reaching the distal small bowel in the 3 cases provided histologic confirmation of the capsule findings. One major although rare limitation of capsule endoscopy is the risk of retention, particularly in patients with RCDII who are particularly at risk of strictures ( Fig. 2 ). Prior radiologic imaging of the small bowel is recommended to rule out stricturing disease and also diagnose extraintestinal lymphoma involvement of rarely giant cavitating lymph nodes ( Fig. 3 ). The second limitation is the need for biopsy during endoscopy ( Fig. 4 ) for definitive diagnosis. In RCDI, histologic examination is similar to that found in active celiac disease with villous atrophy and increased normal IEL. No other diagnostic criteria have yet been defined for RCDI. In contrast, the hallmark abnormal IEL findings, detected using 3 combined techniques, make the diagnosis of RCDII more specific: more than 25% of the CD103+ or CD45 + IELs lacking surface CD3–T-cell receptor (TCR) complexes on flow cytometry ( Fig. 5 ) or more than 50% IELs expressing intracellular CD3ε but not CD8 in formalin-fixed sections ( Fig. 6 ), and/or the presence of a detectable clonal rearrangement of the gamma chain of the TCR in duodenal biopsies ( Fig. 7 ). Similar features allow detection of lymphocytic gastritis and colitis containing the same abnormal population in around 50% and 30% of patients with RCDII, respectively. RCDII may be misdiagnosed when fluorescence-activated cell sorter analysis of freshly isolated IEL is lacking. Discrepancies in diagnostic tools are probably responsible for differences observed between European and North American countries. Heterogeneity in detection of the clonal TCR rearrangement may also explain diagnostic differences. The authors showed that analysis of the delta chain rearrangement may be useful in patients with RCDII presenting with oligoclonal rearrangement of the gamma chain. Other investigators have recently shown interest in detecting the beta chain of the TCR. Nevertheless, evidence that abnormal IELs originate from immature T cells tempers interest in studying beta rearrangement, which occurs later in the ontogeny. Finally, specificity of the PCR product must be shown through formation of homoduplexes.
Diagnosis of RCD
Definition
RCD can be defined as persistent or recurrent severe enteropathy and symptomatic malabsorption in patients with previously confirmed celiac disease who have been on a GFD for at least 12 months (see exception mentioned later) and whose enteropathy or symptoms are not otherwise explained by other conditions or gluten conatamination. A very small subgroup of patients with celiac disease develop primary or secondary resistance to a GFD. Incomplete exclusion of dietary gluten must first be eliminated, because as many as 50% of patients are less than optimally compliant. Persistent or recurrent symptoms of malabsorption with intestinal villous atrophy for at least 12 months of a strict GFD characterize RCD. Diagnosis of this condition is made after other small bowel diseases are excluded, such as autoimmune enteropathy, tropical sprue, or common variable immunodeficiency, and the other conditions that are highly likely to explain the symptoms ( Fig. 1 ). RCD has been subdivided into 2 subgroups: type I (RCDI), defined by persisting villous atrophy despite a strict GFD associated with an increased number of IELs bearing a normal phenotype with surface CD3 and CD8 expression, and type II (RCDII) characterized by clonal expansion of abnormal IELs lacking surface markers CD3, CD8, and T-cell receptors, and preserved expression of intracellular CD3.
Epidemiology
Frequency of RCDI and RCDII remain unknown. In the Derby cohort, West and Holmes report that 0.7% of 713 patients with celiac disease had RCDII. In this latest study, diagnosis of patients with RCDII was based only on aspects of ulcerative jejunitis, which could have either caused an underestimation of RCDII from lack of application of molecular techniques or overestimation from confounding with RCDI. Incidence of T-cell lymphoma associated with celiac disease has been estimated at 0.22 to 1.9 per 100,000 inhabitants per year, although some of these cases may proceed directly from celiac disease, often undetected, to lymphoma without passing through a clinically apparent syndrome of RCD.
Predisposing Factors
Whether patients with RCD have a particular genetic background differentiating them from those with uncomplicated CD is still debated. The small numbers of patients with RCD have had limited genetic investigations. However, severity of celiac disease was correlated with the number of HLA-DQ2 copies: homozygosity for HLA-DQ2 was observed in 25.5% of patients with RCDI, 44.1% of patients with RCDII, and 53.3% of patients with enteropathy-associated T-cell lymphoma (EATL), versus 20.7% of patients with uncomplicated celiac disease and 2.1% of controls. Other genes involved in lymphocyte signaling [genes: SH2B3 (12q34), PTPN2 (18q11), RGS1 (1q31)] are associated with celiac disease and could be involved in the risk of developing overt lymphoma. Ongoing genomewide association studies suggest that the known celiac susceptibility variants may be not found in RCDII.
Another important environmental factor is probably the exposure to gluten. Risk of lymphomatous complications was reported to be 4 times higher in patients not observing a GFD than in those who were compliant. The amount of gluten consumption could be responsible for the differences in terms of severity of celiac disease. A recent study shows that the more severe outcome of celiac disease in southern compared with northern Europe is related to a higher gluten intake.
Diagnosis
Diagnosis of RCD relies on persisting malabsorption and villous atrophy after 1 year of following a strict GFD typically confirmed by expert dietician review. An exception may be necessary for patients who present with severe malabsorption syndrome with ulcerative jejunitis who have strictly adhered to a GFD and are still experiencing an inexorable decline. These cases must be differentiated from the very severe presentation that can occasionally occur in regular celiac disease. Initial endoscopic assessment includes standard or extended upper gastrointestinal endoscopy with biopsy. Sufficient biopsies must be taken to allow for a full assessment of the degree of mucosal injury and for special studies to determine the type of RCD or alternative diagnoses. Careful visual assessment for focal abnormalities that could indicate a prevalent malignancy is essential. Most will be situated in the proximal small intestine but may be beyond the reach of the standard or even an extended endoscopy. Double-balloon enteroscopy is a superior tool for identifying suspicious lesions for EATL or for a better assessment of ulcers, particularly for evidence of ulcerative jejunitis found in roughly 70% of patients with RCDII. Capsule endoscopy is useful for showing the extent of lesions. Capsule endoscopy has superior sensitivity for predicting persistent villous atrophy compared with standard video endoscopy. Furthermore, capsule endoscopy allows the visualization of ulcers all along the intestinal tract, which may suggest RCDII before diagnostic confirmation through direct biopsy sampling. Moreover, the authors diagnosed 3 cases of overt lymphoma when capsule endoscopy showed very suspicious intestinal strictures and jejunal ulcers. Double-balloon enteroscopy reaching the distal small bowel in the 3 cases provided histologic confirmation of the capsule findings. One major although rare limitation of capsule endoscopy is the risk of retention, particularly in patients with RCDII who are particularly at risk of strictures ( Fig. 2 ). Prior radiologic imaging of the small bowel is recommended to rule out stricturing disease and also diagnose extraintestinal lymphoma involvement of rarely giant cavitating lymph nodes ( Fig. 3 ). The second limitation is the need for biopsy during endoscopy ( Fig. 4 ) for definitive diagnosis. In RCDI, histologic examination is similar to that found in active celiac disease with villous atrophy and increased normal IEL. No other diagnostic criteria have yet been defined for RCDI. In contrast, the hallmark abnormal IEL findings, detected using 3 combined techniques, make the diagnosis of RCDII more specific: more than 25% of the CD103+ or CD45 + IELs lacking surface CD3–T-cell receptor (TCR) complexes on flow cytometry ( Fig. 5 ) or more than 50% IELs expressing intracellular CD3ε but not CD8 in formalin-fixed sections ( Fig. 6 ), and/or the presence of a detectable clonal rearrangement of the gamma chain of the TCR in duodenal biopsies ( Fig. 7 ). Similar features allow detection of lymphocytic gastritis and colitis containing the same abnormal population in around 50% and 30% of patients with RCDII, respectively. RCDII may be misdiagnosed when fluorescence-activated cell sorter analysis of freshly isolated IEL is lacking. Discrepancies in diagnostic tools are probably responsible for differences observed between European and North American countries. Heterogeneity in detection of the clonal TCR rearrangement may also explain diagnostic differences. The authors showed that analysis of the delta chain rearrangement may be useful in patients with RCDII presenting with oligoclonal rearrangement of the gamma chain. Other investigators have recently shown interest in detecting the beta chain of the TCR. Nevertheless, evidence that abnormal IELs originate from immature T cells tempers interest in studying beta rearrangement, which occurs later in the ontogeny. Finally, specificity of the PCR product must be shown through formation of homoduplexes.
