Review

Geliac disease, or celiac sprue, is defined as an enteropathy caused by heightened immunologic responsiveness to gluten in genetically susceptible individuals.¹ In the past, celiac sprue was thought of as primarily a gastrointestinal disorder. However, the advent of highly sensitive and highly specific serologic tests over the past decade has led to the increased diagnosis of celiac sprue in adults with predominantly extraintestinal manifestations including osteoporosis, infertility, and neurologic disorders. The prevalence of neurologic disorders including ataxia, peripheral neuropathy, seizures, dementia, myopathy, headache syndromes, anxiety, and depression in established cases of celiac sprue is estimated to range from 6–10%. Ataxia is the most common neurologic manifestation of celiac sprue, and progressive gait and limb ataxia may be the sole manifestation of disease in some patients.² These abnormalities are thought to result from immunologic damage to the cerebellum, posterior columns of the spinal cord, or peripheral nerves, possibly mediated by antigliadin antibodies (AGA), and the term gluten ataxia has been proposed to describe this disorder.³

Maltz and Smith describe a 53-year-old man as having gluten ataxia.⁴ The patient presented with sudden onset ataxia and underwent an extensive battery of laboratory, neurologic, and radiologic investigations, all of which were negative, before finally undergoing celiac serologic testing, which revealed elevated immunoglobulin G (IgG) AGA levels. The patient's immunoglobulin A (IgA) AGA and IgA antiendomysial antibody (EMA) levels were negative, as were his duodenal biopsies.

However, gluten ataxia remains a questionable label, given that only 30% of patients with gluten ataxia (defined as sporadic ataxia with circulating AGA in the absence of any other cause of ataxia) have histologic evidence of an enteropathy diagnostic of celiac sprue.^2,5 It is unclear whether the presence of AGA in patients with gluten ataxia and no other evidence of celiac sprue contributes to the pathogenesis of the ataxia or is merely an epiphenomenon. Supporting the epiphenomenon theory is the fact that 30% of the normal population is AGA-positive and that any condition that causes a breach of the gastrointestinal mucosal barrier such as acid reflux, nonsteroidal anti-inflammatory drug gastritis, gastroenteritis, and Crohn's disease can cause a positive AGA.⁶ In other words, the false-positive rate of AGA is so high that approximately 30% of all patients with ataxia can be expected to have positive AGA serology. Furthermore, despite the presence of AGA, the vast majority of patients labelled as having gluten ataxia fail to respond to the introduction of a gluten-free diet. Whether this represents irreversible gluten-mediated neurologic disease or a lack of cause and effect is unclear.

In contrast to the epiphenomenon theory, recent research by Hadjivasiliou and colleagues² suggests that gluten sensitivity may directly contribute to ataxia in patients with positive AGA but no evidence of enteropathy. Hadjivasiliou and colleagues studied 9 patients labelled with gluten ataxia and 7 disease controls comprised of patients with other causes of ataxia. Whereas all 9 patients with gluten ataxia were AGA-positive, 4 were EMA/tissue transglutaminase (tTG)-positive and had enteropathy, whereas 5 patients were EMA/tTG-negative and had no enteropathy. Hadjivasiliou and associates demonstrated the presence of novel autoantibodies against type 2 tissue transglutaminase (TG2) in the jejunum of all 9 gluten ataxia patients but none in the jejunum of the 7 disease control patients. Intriguingly, the authors were also able to demonstrate the presence of TG2 antibody staining in the autopsy brain tissue of 1 gluten ataxia patient who, despite being EMA/tTG-negative and having a normal duodenal biopsy, had widespread IgA TG2 deposition around vessels in the cerebellum, pons, and medulla. One disease control patient did not have TG2 staining in the autopsy brain tissue. Though these findings are interesting, the number of study participants is small and the brain findings of only 1 patient with gluten ataxia was reported. Ultimately, these findings need to be reproduced in a larger cohort of patients with gluten ataxia before TG2 antibodies can be adopted as a potential diagnostic test for ataxic patients with positive AGA, negative EMA/tTG, and negative duodenal histology.

Although we welcome the case report by Maltz and Smith, it does raise a number of issues. First, the patient did not undergo tTG antibody testing, which is a far more sensitive test for gluten-sensitive enteropathy than AGA. Second, it is unclear how many duodenal biopsies were taken and whether the patient may have been on an empiric gluten-free diet when he was biopsied. Third, patients who are IgG AGA-positive but IgA EMA-negative and IgA AGA-negative, such as this patient, may have IgA deficiency, which is present in 2–10% of the celiac sprue population and which could account for falsenegative serologic results in biopsy-proven celiac sprue. Furthermore, given that nearly 99% of individuals with celiac sprue are either human leukocyte antigen (HLA)-DQ2-positive or HLA-DQ8-positive^7–9 and absence of these alleles virtually excludes celiac sprue, it would be interesting to know the HLA status of this ataxic patient. Finally, the authors do not mention the effect of glutenfree diet on this patient's ataxia. Although the benefits of gluten-free diet in patients with celiac sprue, dermatitis herpetiformis, and a wide range of extraintestinal manifestations of diseases ranging from infertility to bone disease, have long been established,^7,10 unfortunately, the effects of gluten withdrawal on neurologic dysfunction in the setting of celiac sprue has been disappointing. Only one relatively large study reported a good response to strict gluten-free diet in patients with gluten ataxia, even in the absence of enteropathy.¹⁰ It appears that adherence to a strict gluten-free diet was required in patients whose ataxia improved, and even then the clinical response was slower than typically seen with classic celiac sprue. In a number of patients, gluten withdrawal resulted in only a stabilization of the neurologic manifestations rather than a distinct clinical improvement.

In conclusion, gluten ataxia remains as enigmatic as ever and the discovery of novel autoantibodies such as TG2 may help unravel and potentially reverse this neurologic condition. Given the poor sensitivity of AGA and the fact that tTG and EMA, with their far superior sensitivity and specificity, have largely replaced AGA in clinical practice, it is perhaps time we consider distinguishing ataxia in the setting of biopsy-proven celiac sprue as true gluten ataxia, whereas ataxia in the setting of AGA but in the absence of biopsy-proven celiac sprue remains just that—AGA-positive ataxia.

Additional Questions and Answers

Drs. Maltz and Smith posit additional questions for response from the reviewers.

How can a busy gastroenterologist most efficiently distinguish between irritable bowel syndrome (IBS) and celiac disease based on clinical history, especially if a patient meets multiple Rome criteria and invasive studies are not clearly indicated?

On the basis of clinical history, it is not easy to differentiate between celiac sprue and IBS because there is significant overlap between the symptoms of these conditions, even if the Rome II criteria for IBS are strictly applied. Diarrhea in the setting of celiac sprue is often similar to that of IBS patients, ie, postprandial, chronic, intermittent, small volume or watery, and approximately 10% of patients can have constipation. The majority of adults with celiac sprue are either asymptomatic or have extraintestinal manifestations. However, where present, the symptoms of malabsorption, such as nocturnal diarrhea or steatorrhea, weight loss, anaemia, bone pain, family history of celiac sprue, and the presence of other autoimmune disorders should favour the diagnosis of celiac sprue, whereas the absence of these symptoms and the presence of Rome II criteria should favour a diagnosis of IBS. In one large study, over 5% of patients who were referred to secondary care and who fulfilled Rome II criteria for IBS were subsequently found to have celiac sprue.¹¹ The American Gastroenterological Association's guidelines recommend that all patients who meet Rome II criteria for IBS should undergo serologic testing for celiac sprue, and those who are tTG- or EMA-positive should undergo duodenal biopsy.

What is the role of HLA-DQ2 and HLA-DQ8 genetic marker testing, which may occur in the majority of celiac patients but may also occur in up to 30% of patients without the disease?

HLA-DQ2 and HLA-DQ8 tests are tests of exclusion, not of diagnosis. In clinical practice, HLA testing can be useful when proving that a patient does not have celiac sprue. For instance, a clinician suspects that a patient has IBS, and not celiac sprue, on the basis of normal tTG/EMA serology and normal duodenal biopsies, or a patient with IBS goes on an empiric gluten-free diet resulting in negative serology and biopsies, but is reluctant to undergo a gluten challenge. In these settings, negative HLA-DQ2 and HLA-DQ8 results virtually exclude the diagnosis of celiac sprue.

What is the risk of non-Hodgkin lymphoma (NHL) in patients with positive serologies but normal bowel mucosa and no gastrointestinal symptoms? How will gluten-free dietary restrictions affect that risk?

There is a significantly increased risk of NHL in patients with symptomatic celiac sprue but much less than previously thought.¹² The risk of NHL in patients with potential sprue, ie, positive tTG/EMA serology but normal duodenal biopsies and no gastrointestinal symptoms, is unknown. There are some epidemiologic data to suggest that patients with silent celiac sprue have a lower risk of NHL than patients with classic celiac sprue.^12,13 We have no data on the role of gluten-free diet in individuals with potential celiac sprue or silent celiac sprue.

What is the association between celiac sprue and infertility?

There is a strong association between celiac sprue and both female and male infertility. Delayed menarche, early menopause, and poor pregnancy outcomes are welldescribed in patients with celiac sprue, and all women with recurrent miscarriages should undergo serologic testing.^14–16 In men, celiac sprue can cause hypogonadism, immature secondary sex characteristics, and reduced semen quality.¹⁶ Gluten withdrawal and the correction of nutritional deficiencies can lead to a return of fertility in women and men,¹⁶ and there are numerous reports of women investigated for infertility who became pregnant within months of commencing a gluten-free diet after being diagnosed with celiac sprue.

What is the role of the gastroenterologist in dietary advice and counseling?

All patients newly diagnosed with celiac sprue should be referred by their gastroenterologist to a dietician before embarking on a lifelong gluten-free diet. Patients should be counselled about the importance of a lifelong glutenfree diet in significantly reducing their risk of lymphoma, bone disease, infertility and, possibly, associated autoimmune diseases. All patients should be referred to their local or national celiac patient support groups; in addition, websites such as www.celiac.com can prove invaluable by providing regularly updated lists of gluten-free medications and supplements. Patients with established celiac sprue whose symptoms start recurring should also see a dietician because the most common cause of relapse in this setting is inadvertent gluten exposure rather than refractory sprue.

Is periodic antibody testing necessary in patients who have had symptom resolution on a gluten-free diet?

Periodic, usually annual, antibody testing with tTG has replaced periodic AGA testing, as well as periodic invasive and expensive duodenal biopsies, especially in individuals who have symptom resolution on a gluten-free diet. Patients and doctors may find it helpful to remember “the rule of twos,” which states that following a gluten-free diet, symptoms should improve within 2 weeks, serology should normalize within 2 months, and biopsies should normalize within 2 years. Annual antibody testing not only helps monitor dietary compliance but also detects patients who may be ingesting gluten inadvertently.¹⁷ Patients with persistently raised tTG levels despite complying with a gluten-free diet should undergo repeat duodenal biopsies.

At what point is osteoporosis screening indicated in celiac disease patients?

Reduced bone density is the most common complication of celiac sprue, and its prevalence increases with age at diagnosis. Over 70% of patients with untreated celiac sprue have osteopenia,¹⁸ and osteoporosis occurs in over a quarter of all patients.¹⁹ Screening for osteopenia and osteoporosis by dual energy x-ray absorptiometry (DEXA) scanning is recommended at the time of diagnosis for all patients with celiac sprue,^20,21 with repeat DEXA scanning recommended every 2–3 years in patients with abnormal DEXA scans at the outset.

When is endoscopy and biopsy clearly indicated? What is the recommended algorithmic approach to the patient found with positive serologies?

Despite the availability of highly sensitive and specific serologic tests such as tTG and EMA, duodenal biopsies remain the cornerstone of diagnosis. Upper endoscopy and duodenal biopsies are clearly indicated in any individual with positive tTG and/or EMA serology. In general, patients with positive serologies should not automatically be diagnosed with celiac disease and told to start gluten-free diets before they have had duodenal biopsies to confirm the diagnosis. Where the index of suspicion is low and there is no other indication for upper endoscopy (eg, a young patient who meets Rome II criteria but has no family history of celiac sprue and does not have anemia, steatorrhea, or weight loss), then negative tTG/EMA serology should suffice in excluding celiac sprue. Where the index of suspicion is high (eg, in a patient with positive family history, iron- or folatedeficiency anemia, steatorrhea, mouth ulcers, itchy rashes involving extensor surfaces, or associated autoimmune diseases), both tTG/EMA serology and duodenal biopsies should be performed.