Diagnostic Dilemmas in Celiac Disease

Michael X Ma; Mina John; Geoffrey M Forbes

Disclosures

Expert Rev Gastroenterol Hepatol. 2013;7(7):643-655. 

In This Article

Diagnostic Dilemmas in Established CD

The majority of patients with CD will experience clinical improvement following commencement of a GFD. However, ~5% of patients will have nonresponsive CD, defined as persistence of symptoms or serologic and/or histologic abnormalities after 12 months on a GFD.[61] Failure to improve may be due to:

  • Nonadherence to a GFD:

    • Incorrect diagnosis of CD;

    • Refractory CD;

    • Ulcerative jejunitis (UJ) or intestinal lymphoma.

  • Concurrent gastrointestinal disorders.

Nonadherence to a GFD

Approximately 70% of patients will experience some clinical improvement 2 weeks after commencing a GFD, although the rate of response varies.[104] Continued dietary exposure to gluten, often unintentional, is the most common cause of persistent symptoms.[62] On a population level, adherence to GFD is extremely variable with 42–91% of patients estimated to be adherent to the diet, with poorest adherence seen among ethnic minorities and those diagnosed in childhood.[63] Common causes for unintentional gluten exposure include inadequate knowledge of gluten containing products, poor labeling of foods, cross-contamination with gluten containing foods in home cooking and frequent restaurant meals.[61]

Persisting gluten exposure can be determined following detailed history and food diary. However, it may be difficult to completely avoid dietary gluten, as trace amounts can be contained in products labeled as gluten free and regulation of such labeling varies by country. For example, strict food labeling laws in Australia and New Zealand currently define gluten free as 'no detectable gluten' and 'no ingredient derived from oats or malt', with 3 ppm being the limit of gluten detection.[102,105] The current international standard for gluten free foods is ≤20 ppm, set in 2008 by the Codex Alimentarius Commission.[106] This definition was adopted by the European Union in 2009 and is also the standard set by the US Food and Drug Authority. Therefore, foods labeled 'gluten free' in the USA or some European countries may not qualify for such labeling in Australia and New Zealand. Furthermore, patient guidelines may recommend foodstuffs as gluten free based solely on listed ingredient, rather than on formal testing, which may provide an alternative mechanism for inadvertent gluten ingestion.

However, absolute gluten avoidance may not be necessary. In an earlier study evaluating occult gluten intake from grain contaminants among 76 CD patients on a GFD, gluten contamination of 100 ppm, up to a total of 30 mg/day, did not cause histological damage.[64] In contrast, in a prospective randomized controlled trial of 49 biopsy proven CD patients, intake of 50 mg gluten per day for 90 days produced significant damage in the architecture of the small intestine.[14] In practice, the adoption of strict food labeling and dietary adherence is important as it is easy for patients to inadvertently reach gluten thresholds – 50 mg of gluten is ingested in just 3 out of 100 of a single slice of bread. Current evidence based upon population and systematic review data suggests adoption of a single gluten threshold of 20 ppm for gluten contamination,[65] and total gluten intake of upto 10 mg/day as appropriate for most patients with CD,[66] although some patients require levels of gluten intake approaching 0 ppm.[67]

Adherence to a GFD may be assessed by several methods, including regular follow-up with the dietician and/or doctor, serology, small intestinal biopsy and structured surveys.[47] Removal of dietary gluten results in a gradual decline in serum celiac antibody levels, typically normalizing after 3–12 months. Persistently, positive serology 1 year after starting a GFD strongly suggests gluten contamination.[62] Although serology can be used to monitor a patient's adherence, minor transgressions in gluten intake are often unable to be detected and uncertainties remain regarding the optimal timing of testing after commencement of a GFD.[13] Furthermore, seroconversion after GFD also does not necessarily imply mucosal healing, as reflected in a study showing 44% of CD patients with normal anti-tTG IgA after at least 1 year on a GFD had persisting villous atrophy.[68] Mucosal healing, particularly of the proximal small bowel, typically occurs later than clinical and serological improvement, taking up to months or years to normalize in some cases.[69] As a lack of healing may be associated with increased risk of lymphoma, bone disease and development of refractory CD, some guidelines suggest performing a follow-up biopsy in adults 2 years after starting a GFD.[47]

Refractory CD

Refractory CD (RCD) is a rare complication, defined as persistence of malabsorptive symptoms and villous atrophy for >6 months despite strict adherence to a GFD and negative celiac serology.[7,70] Diagnosis of RCD is important as specific treatments (glucocorticoids and immunosuppressives) exist. Clinically, RCD is suggested by presence of severe malnutrition (mean body mass index <18), protein losing enteropathy (90%) and endoscopic features of UJ (70%).[70]

Two types of RCD are recognized,[71] diagnosed pathologically by analysis of IEL phenotype based on flow cytometry of cell preparations of fresh duodenal biopsies[72] or immunohistochemical staining of paraffin-embedded sections.[73] In RCD1, the IELs have a normal phenotype (predominantly oligoclonal CD3+/CD8+ T cells), whereas RCDII is characterized by a clonal expansion of IELs lacking surface CD3 and generally CD8 but containing intracellular CD3ε.[70]

For RCD diagnosis, the advantage of flow-based analysis is the ability to quantitate IELs based on surface CD3/8 expression and intracellular CD3; however, there may be technical issues associated with the yield of lymphocytes from biopsy material as well as contamination with lamina propria lymphocytes. Conversely, immunohistochemical staining ensures visual localization of cells to the epithelium but cannot determine aberrant phenotype based on intracellular CD3 expression. The diagnostic cutoffs for proportions of aberrant IELs also differ based on technique, being >20% by flow-based methods[74] and >40% by immunohistochemistry.[73]

UJ & Lymphoma

UJ and enteropathy-associated T-cell lymphoma (EATL) are rare complications of CD and should be considered in patients with RCD unresponsive to treatment. Both have similar clinical features and share a common pathogenesis via aberrant T-cell monoclonality.[75] UJ is characterized by the presence of multiple chronic, benign appearing ulcers occurring most often in the jejunum, and may be complicated by luminal stricturing and recurrent bowel obstruction. Multifocal ulcerating, proximal small intestinal nodules are features of EATL. In both conditions, suggestive clinical features include recurrence of abdominal pain, weight loss and diarrhea despite adherence to a GFD. Fevers and night sweats point toward lymphoma.[16]

Suggested initial investigations include abdominal imaging (computed tomography or magnetic resonance enterography) plus gastroscopy. Capsule endoscopy is sensitive in demonstrating small-bowel pathology and can be considered if gastroscopy is normal. If EATL is suspected, additional investigations including positron emission tomography scan and enterography for histological sampling can aid making a diagnosis. Full-thickness surgical intestinal biopsy may be required if radiological and endoscopic testing is inconclusive.

Concurrent Gastrointestinal Disorders

A number of conditions are associated with CD and may cause persistent symptoms despite adherence to a GFD. These conditions include:

  • IBS

  • Lactose or fructose intolerance

  • IBD

  • Microscopic colitis

  • Exocrine pancreatic insufficiency

  • Small-bowel bacterial overgrowth.

The association of IBS with CD is well recognized.[16,39,76] A recent meta-analysis reports symptoms compatible with IBS are more common in those with CD than the general population, occurring in almost 40% of patients with CD.[76] A shared pathophysiology is possible as both visceral hypersensitivity and dysmotility have been reported in those with IBS and CD. In the same study, there was also a trend toward a greater prevalence of IBS-like symptoms in those who were not strictly adherent to a GFD, compared with adherent patients, suggesting that gluten contamination may contribute to IBS-type symptoms in a proportion of patients with CD.[76] However, as both conditions share significant symptom similarity, investigation for other causes of nonresponsive CD is strongly advised prior to concluding IBS as the cause of persistent symptoms. Lactose or fructose intolerance, as a result of small intestinal mucosal damage from CD, may be a cause of ongoing diarrhea, flatulence and abdominal pain. In a study of 113 CD patients with persistent symptoms after 6 months on a GFD, 9 were found to have lactose intolerance by hydrogen breath tests.[77] Recognizing that CD and lactose intolerance can co-exist, exclusion of dietary produce has sometimes been recommended in the first 6 months after commencing a GFD to allow recovery of brush border disaccharidase activity.[78] IBD may also occur concurrently with CD; a retrospective review of 455 patients with CD found an age- and sex-adjusted prevalence rate ratio for ulcerative colitis of 3.56 (95% CI: 1.48–8.56) and Crohn's disease of 8.49 (95% CI: 3.53–20.42).[79]

Microscopic colitis is associated with ~4% of patients with CD. It should be considered when diarrhea is the predominant symptom[80,81] and is important to identify as specific treatments are available.[61] In a separate study of 66 patients with CD and chronic diarrhea, 20 were found to have concomitant pancreatic insufficiency based upon low fecal elastase-1 levels.[82] Diarrheal symptoms improved in 18 of these patients after commencing pancreatic enzyme supplementation. A link between small-bowel bacterial overgrowth and CD has also been suggested, though two small therapeutic studies of rifaximin yielded conflicting results.[83,84]

processing....