Introduction Historically Brunner's glands were thought to interfere with interpretation of duodenal bulb biopsies. Recent reports suggest that the duodenal bulb may be the only site to demonstrate villous atrophy (VA) in coeliac disease (CD). There are few data on the prevalence of lesions in non-coeliac patients.

Methods We aimed to compare the histological findings in the duodenal bulb and distal duodenum of patients with CD against controls having gastroscopy with duodenal biopsy. Indications included positive coeliac serology, family history, diarrhoea, and iron deficiency anaemia. A total of 215 patients were prospectively accrued. Biopsies were graded using the Marsh criteria by a single pathologist.

Results 215 patients, (150 female, 65 male) median age 48 years (range 16-89) were analysed. 17 patients demonstrated VA in the bulb only. 1 had VA in distal duodenum only. 1 control patient with HIV enteropathy had VA in the bulb and distal duodenum. Both new diagnosis CD (8%, p<0.05) and CD in remission (15%, p<0.001) were more likely than controls to have VA in the bulb alone. When comparing the Marsh grade score for the bulb and the duodenum, 24/59 of patients in remission had a discrepancy of at least 1 Marsh grade compared with only 8/91 new diagnosis CD (p<0.0001), and 4/65 controls (p<0.0001). 21/24 of the patients in remission with a discrepancy had the more severe lesion in the bulb (Abstract 087).

Conclusion Villous atrophy may be present only in the duodenal bulb. In this study 7/91 (8%) of new diagnosis CD and 9/59 (15%) of CD in remission demonstrated VA in the bulb alone. When looking at CD in remission, 41% of patients have a discrepancy in severity of the histological lesion between the bulb and the duodenum; in this group the most severe lesion is nearly always in the bulb. This is the first study to demonstrate that optimal assessment of both new diagnosis CD and those in remission requires a duodenal bulb biopsy in addition to distal duodenal biopsies.