Celiac Disease-Even a Neurological Disorder

Does Cryptic Gluten Sensitivity Play a Part in Neurological Illness? Hadjivassiliou M, Gibson A, Davies-Jones GAB, Lobo AJ, Stephenson TJ, Mitford-Ward A. Lancet 1996;347:369-71. Summary: This article describes the occurrence of adult celiac disease in neurological disorders. The authors utilized antigliadin antibody measurement as a screening method for celiac disease among 147 patients having various neurological disorders. Fifty-three patients had neurological disease of unknown cause (group 1), and 94 had a specific neurological diagnosis (group 2). Fifty healthy blood donors constituted the control group (group 3). Positive titers for antigliadin antibodies were found in 30 (57%) patients in group 1, five (5%) in group 2, and six (12%) in group 3. All patients with positive antigliadin antibodies in groups 1 and 2 were asked to undergo duodenal biopsy; four and two refused, respectively. Nine (16%) patients in group 1 were found to have villous atrophy compatible with celiac disease, and in addition 10 (18%) had nonspecific, minor changes. In group 2, all three duodenal biopsy specimens were normal. Only one patient with celiac disease had anemia; the other eight had no symptoms or signs of malabsorption. The authors recommend screening for celiac disease with antigliadin antibodies in all patients with neurological dysfunction of unknown cause. Given the high frequency of antigliadin antibodies, they further hypothesize that these antibodies could be directly neurotoxic. Comment: Today celiac disease has been detected to an increasing extent by serological screening among various risk groups, such as first-degree relatives of celiac patients (Lancet 1991;338:1350-3) and patients with insulin-dependent diabetes mellitus (Arch Dis Child 1984;59:739-42; Scand J Gastroenterol 1989;24:81-4), autoimmune thyroid diseases (Eur J Endocrinol 1994;130:137-40), and alopecia areata (Gastroenterology 1995;109:1333-7). Most of the screened patients with celiac disease did not suffer from overt malabsorption, and many were entirely asymptomatic and therefore had silent celiac disease. The current article maintains that neurological dysfunction of unknown cause is a novel indication for serological screening for celiac disease. Moreover, the results of Hadjivassiliou et al. nicely exemplify the overall changing pattern of celiac disease. According to earlier studies, neurological complications have not been uncommon in celiac disease (Brain 1966;89:683-722). Such complications have been reported in as many as 37% of patients with celiac disease, but these were usually caused directly by the consequences of malabsorption, for example deficiency of vitamin D or hypocalcemia (Neurol Sci 1971;14:125-41). At present celiac disease is detected earlier, and severe malabsorption has therefore become rare. By contrast, celiac disease without any symptoms or signs of malabsorption can be found among patients suffering from neurological dysfunctions of otherwise unknown causes. As Hadjivassiliou et al. have shown, these patients can be found by serological screening. We have previously described five patients with celiac disease, brain atrophy, and dementia (Neurology 1991;41:372-5). In contrast to the results of Hadjivassiliou et al., four of these five patients had unequivocal signs of malabsorption, which suggests that their disease had remained undiagnosed for a long time. Nevertheless, both the findings of Hadjivassiliou et al. and our own raise the question whether ingested gliadin itself could be neurotoxic in the case of patients with gluten-sensitive disease. Even though the pathogenesis of celiac disease remains unknown, immunological mechanisms in genetically susceptible individuals are involved. These same mechanisms could contribute to the development of nervous system damage. The authors hypothesize that antigliadin antibodies could be neurotoxic, since they were detected with such high frequency. On the other hand, we believe that antigliadin antibodies are not necessarily the most significant antibodies in celiac disease. First, a high frequency of antigliadin antibodies is not an uncommon finding when asymptomatic patients are screened. The authors themselves detected these antibodies in 12% of healthy control subjects. The positive predictive value of IgA class antigliadin antibodies was 20% when we screened patients with autoimmune thyroid diseases for celiac disease (Eur J Endocrinol 1994;130:137-40). Uibo et al. (Dig Dis Sci 1993;38:2034-7) found 3.5% of 1,461 persons in an Estonian population to be positive for antigliadin antibodies; none of these subjects, however, had small bowel villous atrophy. Second, antigliadin antibody positivity is not closely linked to celiactype genetics (Lancet 1992;338:1350-3), namely, to HLA DQA1*0501, HLA DQB1*0201 alleles. By contrast, patients with antireticulin or antiendomysium antibodies often carry these alleles, and they can also have latent celiac disease. We have earlier shown celiac disease to develop in patients who were positive for antireticulin antibodies, but whose initial duodenal biopsy was normal. This development was much more uncommon among patients positive for antigliadin antibodies only (Gut 1991;32:1412-4; Scand J Gastroenterol 1993;28:595-8). In view of this, our alternative suggestion is that antireticulin or antiendomysium antibodies may be involved in immune mechanisms that lead to celiac disease or even to its neurological complications. In fact, these antibodies are autoantibodies (Lancet 1991;338:724-5), and we have recently shown them to be directed against human fibroblast-derived antigens (Lancet 1993;34:420-3). Hadjivassiliou et al. did not report the occurrence of antireticulin or antiendomysium antibodies among their patients. What seems to us intriguing is that 18% of their patients suffering from neurological dysfunction of unknown cause had minor inflammatory changes in duodenal specimens. Could this mean celiac disease latency? It would be interesting to know whether these patients carry the HLA DQA1*0501 and DQB1*0201 alleles, or whether they are positive for antireticulin and antiendomysium antibodies. Moreover, an increased count of γ/δ T-cell receptor-bearing intraepithelial lymphocytes could mean that these patients actually have a latent form of celiac disease (Gut 1991;32:1412-4; Lancet 1992;339:1500-3). We have one case of latent celiac disease in a patient with neurological disease. This patient has had diabetes mellitus since the age of 3. At the age of 12, he was positive for antireticulin antibodies but had a normal small bowel villous architecture. Fifteen years later, the patient had brain atrophy and dementia, and at this time the duodenal mucosa showed villous atrophy and crypt hyperplasia as well. This might indicate that the ingested gliadin had been harmful to the nervous system even before the enteropathy had developed. Even though the significance of antireticulin, antiendomysium, and antigliadin antibodies in the etiology of neurological dysfunction remains obscure, the findings of Hadjivassiliou et al. are of particular interest. They show that many patients with nonspecific neurological dysfunction suffer from silent celiac disease. Surveillance and further studies will show whether neurological symptoms are alleviated on treatment with a gluten-free diet. We have justified our own policy of screening for silent celiac disease on the basis of the risk of malignancy (Gut 1989;30:333-8) and the development of osteopenia (Scand J Gastroenterol 1994;29:457-61; Gastroenterology 1995;109:122-8) in untreated celiac disease. A risk of neurological complications in silent celiac disease can now be added to our list. Pekka Collin; Markku Mäki Departments of Medicine and Paediatrics; Tampere University Hospital; Medical School and Institute of Medical Technology; University of Tampere, Tampere, Finland