Article Text

Download PDFPDF
OC-081 Can a “risk of coeliac disease” scoring system reduce the number of patients referred for duodenal biopsy?
  1. J S Leeds1,
  2. N Libzo2,
  3. R Sidhu3,
  4. K E Evans3,
  5. M Kurien3,
  6. E J Hall3,
  7. V S Jandu3,
  8. A D Hopper3,
  9. P Basumani4,
  10. K L Dear5,
  11. M E McAlindon3,
  12. D S Sanders3
  1. 1Department of Gastroenterology, Royal Hallamshire Hospital, Sheffield, UK
  2. 2Department of Gastroenterology, Northern General Hospital, Sheffield, UK
  3. 3Department of Gastroenterology and Liver Unit, Royal Hallamshire Hospital, Sheffield, UK
  4. 4Department of Gastroenterology, Rotherham District General Hospital, Rotherham, UK
  5. 5Department of Gastroenterology, Chesterfield Royal Hospital, Chesterfield, UK


Introduction Coeliac disease (CD) affects 1% of the population. Despite this prevalence, the majority of individuals are undetected and patients present with atypical/subtle symptoms contributing to under diagnosis. The aim of this study was to determine the importance of different presenting characteristics and risk factors to provide a biopsy reducing strategy.

Methods Patients referred to the SYLGRG were investigated for CD using antiendomyseal and anti-tissue transglutaminase antibodies plus total IgA level. Demographics, reason for referral and comorbidities were noted. Patients with positive antibodies or IgA deficiency were offered a duodenal biopsy. Duodenal biopsies were graded using the modified Marsh criteria with grade 3 (villous atrophy) taken as diagnostic of CD. Multivariate analysis was performed to identify independent risk factors. Using these factors a coeliac risk score was constructed and tested on a second cohort (n=609).

Results 4089 patients were assessed (mean age 55.8, 2392 females). 386 patients had positive profiles (9.5%, 8.6–10.4%) of which 128 had CD (3.1%, 2.6–3.7%). Univariate analysis showed CD was associated with age <55 years (OR 2.2), IBS symptoms (OR 2.8), anaemia (OR 4.8), diarrhoea (OR 3.6), positive antibodies/need for duodenal biopsy (OR 28.7). Analysis of comorbidities revealed CD was associated with osteoporosis (OR 4.4), autoimmune disease (OR 2.3) and a family history (FHx) of CD (OR 9.0). Linear regression showed that age<55, FHx CD, Anaemia and Osteoporosis were independent risk factors for CD and these factors were then modelled and attributed scores. Age <55 = 1, anaemia=1 and EMA positive=2 giving compositae scores from 0 to 4. Scores of 0–2 corresponded to a very low risk of CD (0.84%, 0.6–1.2) and scores of 3+ corresponded to a risk of 13%–80%. Receiver operating curve analysis for the score gave an area under the curve of 0.79 (0.75–0.85, p<0.001). When applied to the validation cohort, scores of 0–2 corresponded to 0% risk of CD (0.0–0.7) and scores of 3+ corresponded to a risk of 4%–50%.

Conclusion CD accounted for one in 30 referrals. CD was associated with younger age, anaemia, positive family history and osteoporosis. Modelling of these risk factors lead to a potentially valuable scoring system that could be used to determine pre-biopsy risk of CD and in many patients avoid unnecessary biopsy. This score needs to be validated in other cohorts.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.