Article Text
Statistics from Altmetric.com
Editor,—We were interested to read the paper by Kingham and Parker (Gut1998;42:120–2) as we have recently reported a study exploring the association between primary biliary cirrhosis (PBC) and coeliac disease.1
We tested serum samples from 57 patients with biopsy proved PBC for IgA class endomysial antibody. Six (11%) had endomysial antibody though only four agreed to undergo small bowel biopsy, which confirmed coeliac disease in all cases. As we have found endomysial antibody to have 100% specificity for coeliac disease,2 11% is likely to be the true prevalence of coeliac disease among our patients with PBC, which is over 15 times that of the general population in Northern Ireland.3 None of the four patients with biopsy proved coeliac disease showed improvement in liver biochemistry after exclusion of dietary gluten.
Our experience of screening patients with coeliac disease for PBC was less productive.4 Of the 129 patients with coeliac disease screened prospectively for liver disease at the time of diagnosis, none had serum antimitochondrial antibodies. The commonest abnormality in our coeliac patients was a rise in aspartate or alanine aminotransferase, or both, (15%) which was not associated with serum antinuclear or smooth muscle antibodies and which resolved following gluten exclusion.
We agree that patients with PBC should be routinely screened for coeliac disease, though endomysial antibody is likely to be superior to antigliadin antibody for this.2 Apart from the possibility of preventing long term consequences like malignancy and osteoporosis, non-specific symptoms conventionally attributed to PBC, such as fatigue and abdominal pain, may be actually due to coeliac disease and respond to a gluten-free diet.
References
Primary biliary cirrhosis and coeliac disease: a study of relative prevalences
Editor,—We read with interest the study from Swansea (Gut1998;42:120–2) reporting that the prevalence of coeliac disease in patients with primary biliary cirrhosis (PBC) is 6%. The diagnosis of coeliac disease in this group can easily be missed as some of the clinical features may be attributed to cholestasis. This is important in a population where nutrition may be critical, such as those who may eventually undergo liver transplantation for end stage liver disease.
One non-invasive screening strategy is to use the highly specific and sensitive IgA endomysial antibody assay to screen all new patients with PBC, and to perform duodenal biopsies on positive patients for confirmation. We recently investigated whether this yielded occult cases in our population of patients with PBC.
We studied 87 antimitochondrial antibody positive patients with PBC for the presence of endomysial antibody on indirect immunofluorescence. We also included 24 non-PBC cholestatic controls (13 with extrahepatic obstruction, 11 with primary sclerosing cholangitis) and four coeliac serum samples from newly diagnosed patients in a blinded fashion as methodological controls. None of the cholestatic control serum samples showed staining whereas two of the 87 PBC samples stained positively for endomysial antibody: all four coeliac samples were positive.
On review of the records, both of the patients with PBC positive for endomysial antibody had been screened for coeliac disease already on the basis of symptoms alone, and the diagnosis confirmed by duodenal biopsy. None of the remaining 85 patients with PBC had any suggestion of occult coeliac disease. Thus it seems that no new cases were detected by screening with endomysial antibody alone, and the prevalence of coeliac disease in our population was 2.3%. We do not therefore routinely screen all patients with PBC for coeliac disease, but if there is clinical suspicion we measure endomysial antibody in addition to the IgA and IgG gliadin antibodies suggested by Kingham and Parker.
Reply
Editor,—We thank Drs Dickey and McMillan, and Fidler et al for their interest in our paper and note the higher prevalence of coeliac disease among patients with PBC in Northern Ireland but lower in London than that found in Swansea. The precise methods used for case ascertainment in these other populations are not given so the results are not directly comparable with ours. However the very high incidence of coeliac disease in Ireland is well described and a higher than average incidence in the largely Celtic population of South West Wales would not be unexpected. These regional variations in the incidence of coeliac disease in the general population might account for differences in relative prevalences of the two conditions.
If Dickey and McMillan were to repeat their search for PBC among their patients with coeliac disease they should now of course identify those four whom they have shown to have both conditions provided the surveys cover the same patient population.
Whether or not to screen patients with PBC for coeliac disease will depend on clinicians’ perceptions of published figures and their own experience. However, we know that a proportion of coeliac patients have no specific symptoms, so reliance on symptoms alone is inadequate for detecting all cases.
We agree that endomysial antibody now looks to be a more specific marker for coeliac disease than anti-gliadin antibody though the latter assay is more widely available, can be automated and is cheaper.