Abstract

Background: Small bowel Crohn’s disease (CD) can be challenging to diagnose. The Montreal classification suggests that ileal disease will be recognised more frequently with the advent of better small bowel imaging. Isolated small bowel disease has been reported in up-to one third of CD patients. The presence or absence of inflammation in the terminal ileum at ileo-colonoscopy is often considered the gold-standard for CD assessment. Our aim was to evaluate the accuracy of terminal ileal (TI) inflammation/ulceration in predicting active small bowel CD. Methods: A retrospective review was carried out utilising our IBD database of patients referred for investigation of potential active small bowel CD over two years. Patients with both established and suspected CD who had undergone ileocolonoscopy with biopsy and capsule endoscopy (CE) were included. Small bowel CD was defined by either positive histology or CE with clinical response. Exclusion criteria included less than 12 months follow up, a documented history of NSAID exposure and incomplete CE studies. Correlation of TI findings with small bowel Crohn’s disease was performed using SPSS 19. Results: In total, 142 patients were identified, 57% were female with a mean age of 45 (range 19 70) and a mean follow up of 18 months (range 12 36). In all, 62 (44%) had active small bowel CD. TI inflammation/ulceration was seen in 44 (30%) patients and histology correlated with positive TI findings in only 13 (30%) patients. In total, 98 patients had normal TI findings and within this group 30 (30%) had active small bowel CD on CE or follow up. The sensitivity and specificity of TI inflammation/ulceration in diagnosing active small bowel CD are 51% and 85% respectively with a PPV of 71% and a NPV of 30%. Table 1. Of interest, both the PPV and NPV increased to 96% and 71% in the known CD cohort but was not as accurate in patients with suspected CD (PPV 45%, NPV 88%). Of interest, the correlation between visually reported TI inflammation/ ulceration with ultimate histology or diagnosis was very poor with r values of 0.4 and 0.1.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call