Background and aims: Postsurgical recurrence (PSR) is common in patients with Crohn's disease (CD) who have undergone surgery. Endoscopy is crucial for the diagnosis of PSR, showing also high prognostic value. Bowel sonography (BS) is accurate for CD diagnosis, but its role in PSR detection and grading has been poorly investigated. The aim of this study was to evaluate the diagnostic accuracy of BS compared to endoscopy in the detection of PSR.
Materials and methods: Between March 2002 and October 2005, to gain evidence of possible PSR, we prospectively performed endoscopy and BS in 45 CD patients who had undergone previous bowel resection. Endoscopy and BS were carried out 1 year after surgery, with diagnosis and grading of PSR made in accordance with Rutgeerts. BS was considered suggestive for PSR in the presence of bowel wall thickness (BWT)>3 mm. Also, an ROC curve was constructed to define the best cutoff value for BWT to differentiate mild from severe PSR (grade 1-2 vs 3-4 of Rutgeerts).
Results: Of the 45 patients with CD, 24 showed endoscopic evidence of PSR (53%). Severe endoscopic PSR was present in 16 patients (66%). Sensitivity, specificity, and positive and negative predictive values of BS were 79%, 95%, 95%, and 80%, respectively, with a sensitivity of 93% for severe PSR. On the ROC curve, a BWT>5 mm showed sensitivity, specificity, and positive and negative predictive values of 94%, 100%, 100%, and 96%, respectively, in differentiating mild from severe PSR, in remarkable agreement with endoscopy (kappa=0.90).
Conclusions: BS shows good sensitivity and high specificity for the diagnosis of PSR in CD, with a BWT>5 mm being strongly indicative of severe endoscopic PSR. Hence, BS could replace endoscopy for the diagnosis and grading of PSR in patients who comply poorly with the endoscopic examination.