Assessment of disease severity on capsule endoscopy in patients with small bowel villous atrophy

Background and aim: There is a lack of uniformity of reporting on features of celiac disease (CD) on small bowel capsule endoscopy (SBCE). This makes determining extent of disease and comparison of severity of disease challenging.

Methods: De-identified SBCEs of 300 patients (78 CD [26%], 18 serology negative villous atrophy [6%], and 204 controls with normal duodenal histology [68%]) were included. Videos were reviewed by two experts. All patients had duodenal histology taken within 2 weeks of SBCE. The degree of agreement in CD features and extent of disease was then determined. The resulting score for each factor was used to determine overall severity of disease.

Results: There was substantial agreement in the kappa coefficient for the detection of CD features between reviewers (0.67). Agreement for extent of affected small bowel (SB) mucosa was high (0.97). On multiple regression analysis, several features of CD correlated with extent of affected SB mucosa for both reviewers. The odds ratios derived from this analysis were then used to score features of CD, enabling scores of severity to be calculated for each patient. The median overall scores for patients increased significantly according to the independent classification of severity by the capsule reviewers: mild (20, 0-79), moderate (45, 25-123), and severe (89, 65-130) (P = 0.0001).

Conclusion: The good correlation of CD scores between expert reviewers confirms the validity of features of CD on SBCE. An objective score of CD features in the SB is useful in the follow up of patients with CD and serology negative villous atrophy.