Endoscopic Lesions of Postoperative Anastomotic Area in Patients With Crohn's Disease in the Biologic Era: A Japanese Multi-Centre Nationwide Cohort Study.

Background and aims: Many patients have endoscopic evidence of recurrent Crohn's disease [CD] at 1 year after intestinal resection. These lesions predict future clinical recurrence. We endoscopically evaluated postoperative anastomotic lesions in CD patients from a large cohort of postoperative CD patients.

Methods: We retrospectively enrolled CD patients who underwent surgical resection between 2008 and 2013 at 19 inflammatory bowel disease [IBD]-specialist institutions. The initial analyses included patients who underwent ileocolonoscopy ~1 year after intestinal resection. Follow-up analyses assessed any changes in the endoscopic findings over time. We evaluated the postoperative endoscopic findings, which were classified into four categories [no lesion, mild, intermediate, severe] at the sites of the anastomotic line and peri-anastomosis.

Results: In total, 267 CD patients underwent postoperative ileocolonoscopy. Postoperative anastomotic lesions were widely detected in index ileocolonoscopy [61.0%] and were more frequently detected in follow-up ileocolonoscopy [74.9%]. Endoscopic severity also increased. Patients with intermediate or severe peri-anastomotic or anastomotic line lesions at the index ileocolonoscopy required significantly more interventions, including endoscopic dilatation or surgery, than patients with mild lesions or no lesions.

Conclusions: Frequent anastomotic lesions were observed at the postoperative index ileocolonoscopy. These gradually increased for subsequent ileocolonoscopy, even in the biologic era. Regarding lesions on the anastomotic line, intermediate lesions on the anastomotic line [e.g. irregular or deep ulcers] might be considered recurrent disease, and mild lesions [e.g. linear superficial ulcers] might be considered non-recurrent disease. Prospective studies are needed to resolve this issue, including treatment enhancement.