Introduction Coeliac disease is an autoimmune disorder of the small bowel with a prevalence as high as 1:100 in the UK and Ireland. The gold standard for diagnosis is to identify the characteristic histopathological changes (based on the modified Marsh criteria) from an adequate small bowel biopsy. However non-invasive serological blood tests are often the first line investigation. Serological testing is reported to have both high sensitivity and specificity with the sensitivity and specificity of IgA anti-tissue transglutaminase antibodies (tTG) being higher (99% and >90%) than IgA anti-gliadin (46–100% and 86–100%) and IgA anti-endomysium (74–100% and 91–100%). However, in a study of 26 UK patients with coeliac disease Smith-Laing et al (Clinical Medicine 2009) raised the issue of limitations of serological testing reporting discrepancy between histology and anti-tTG in as many as 38.5%. Given our reliance on serological testing the results were of concern. The objective of this retrospective study was to analyse the results of serological tests for coeliac disease in consecutive patients with duodenal biopsies confirming the diagnosis.

Methods Results of duodenal biopsies which fulfilled the histological criteria for coeliac disease between 2005 and 2010 at two UK district general hospitals (King George hospital, Ilford and Queen's hospital, Romford) were correlated with coeliac serological tests. IgA tTG antibodies, IgA anti-endomysium antibodies, and serum IgA levels were recorded. Serological testing done before or within a month of biopsy was noted. Reference range for tTG was >15 U positive.

Results There were 182 positive duodenal biopsies. Serological tests were not performed in 35 patients. Of the remaining 147, sixteen were excluded (IgA not measured—6, IgA deficient—3, serological testing done more than a month after biopsy date—6, other—1). Complete data were therefore available in 131. Nine patients with confirmed coeliac disease had anti-tTG levels below the diagnostic range. Of these, three had positive anti- endomysium antibody. Thus anti-tTG levels suggestive of coeliac disease were found in 122 of 131, demonstrating sensitivity of 0.93 with combined sensitivity of tTG and anti-endomysium of 0.954.

Conclusion This retrospective study reassuringly demonstrates that there is not a significant number of serologically negative but positive biopsy cases. A proportion of patients will not have serological tests going straight to endoscopy as first line investigation for their anaemia. Serological testing remains useful in primary care and for physicians to diagnose coeliac disease; however it is important to be aware of the small number of cases (approximately 5%) that will be missed when relying on serology alone.

Competing interests None declared.