Gluten sensitivity: a many headed hydraBMJ 1999; 318 doi: http://dx.doi.org/10.1136/bmj.318.7200.1710 (Published 26 June 1999) Cite this as: BMJ 1999;318:1710
Heightened responsiveness to gluten is not confined to the gut
- M Hadjivassiliou, Senior registrar in neurology. (, )
- R A Grünewald, Acting head, academic section of clinical neurology.,
- G A B Davies-Jones, Consultant neurologist.
In a lecture entitled “On the coeliac affection”1 given in London in 1887 Dr Samuel Gee first described the condition we now refer to as coeliac disease or gluten sensitive enteropathy. With clinical manifestations confined to the gastrointestinal tract or attributable to malabsorption, it was logical to assume that the key to the pathogenesis of this disease resided in the gut. However, focusing diagnostic criteria on the gut (as most physicians still do) has delayed the appreciation of the wider spectrum of gluten sensitivity.
The treatment of coeliac disease remained empirical until 1940-50, when the Dutch paediatrician Willem Dicke noted the deleterious effect of wheat flour on individuals with coeliac disease.2 Removal of all dietary products containing wheat resulted in complete resolution of the gastrointestinal symptoms and a resumption of normal health. The introduction of the small bowel biopsy in 1950-60 confirmed the gut as the target organ in coeliac disease. The characteristic features of villous flattening, crypt hyperplasia, and increase in intraepithelial lymphocytes with improvement on gluten free diet became the mainstays of the diagnosis of coeliac disease.
However, in 1966 Marks et al showed an enteropathy with a striking similarity to coeliac disease in 9 out of 12 patients with dermatitis herpetiformis.3 The enteropathy and the rash were gluten dependent and the skin disease could occur even without histological evidence of gut involvement. This discovery started to shift the emphasis from the gut as the sole protagonist in this disease. With dermatitis herpetiformis too came the concept of “latent” gluten sensitivity. The term is now used to describe people with a histologically normal small bowel while on a normal diet who at some stage of their lives have had or will have an abnormal small bowel that responds to a gluten free diet.4
Also in 1966 Cooke and Thomas-Smith published a paper on neurological disorders associated with adult coeliac disease.5 Further case reports have since been published, but most are based on patients with coeliac disease who later develop neurological dysfunction, implying that gut disease is a prerequisite. We have, however, shown that neurological dysfunction can not only precede coeliac disease but can also be its only manifestation.6 Of even more interest is the demonstration of a high prevalence of circulating antigliadin antibodies (IgG, IgA, or both) in patients with neurological dysfunction of obscure aetiology (57% v 5% in neurological controls and 12% in normal controls).7 Only 35% of these patients had histological evidence of coeliac disease. The remaining 65% have gluten sensitivity where the target organ is the cerebellum or the peripheral nerves, a situation analogous to that of the skin in dermatitis herpetiformis.
In the light of these findings the specificity of antigliadin antibodies has been questioned yet again. When the histological criteria for coeliac disease are used as the gold standard, IgG antigliadin antibody has low specificity. Nevertheless, IgG antigliadin antibodies have a high sensitivity not only for patients with coeliac disease but also for those with minimal or no bowel damage where the principal target organ is the cerebellum or peripheral nervous system. Supportive evidence for this contention comes from the HLA genotype of patients with neurological disorders associated with gluten sensitivity. As coeliac disease has one of the strongest HLA associations of any immune disease (HLA DQ2 in more than 90% of patients) one would expect that patients positive for antigliadin antibodies and with normal duodenal mucosa should have a similar HLA genotype if they are truly gluten sensitive. In fact 85% of our patients with neurological disorders associated with gluten sensitivity have an HLA genotype in keeping with coeliac disease compared with 25% of the normal population.8
Unlike antiendomysium or antireticulin antibodies, antigliadin antibodies are antibodies against the extrinsic causal factor for gluten sensitivity. Antiendomysium antibodies may be more specific for coeliac disease, but no large scale data are available as yet on their specificity or sensitivity in patients with gluten sensitivity where the immunological target organ may be other than the gut.
The typical clinical expression of a patient with gluten sensitivity where the sole manifestation is neurological is cerebellar ataxia, often with a peripheral neuropathy.9 Most of these patients will have histologically normal mucosa on biopsy and few or no gastrointestinal symptoms. Both the ataxia and the neuropathy may be reversible with adherence to a gluten free diet.9
Marsh's “modern” definition of gluten sensitivity is to be recommended: “a state of heightened immunological responsiveness to ingested gluten in genetically susceptible individuals.”10 Such responsiveness may find expression in organs other than the gut. Gastroenterologists, dermatologists, neurologists, and other physicians need to be aware of these developments if the diagnosis and treatment of the diverse manifestations of gluten sensitivity are to be advanced. The aetiology of such diverse manifestations presents the next challenge.
The authors have received financial support for research from SHS International Ltd and Ultrapharm. MH has been funded by SHS to attend a conference on coeliac disease.