P110. Surveillance for colorectal cancer in colitis patients: effect of the implementation of new British and American guidelines on neoplasia yield
E. Mooiweer1, A.E. van der Meulen2, A.A. van Bodegraven3, J.M. Jansen4, N. Mahmmod5, J. Nijsten1, M.G. van Oijen1, P.D. Siersema1, B. Oldenburg1, 1University Medical Centre Utrecht, Department of Gastroenterology and Hepatology, Utrecht, Netherlands, 2Leiden University Medical Centre, Department of Gastroenterology and Hepatology, Leiden, Netherlands, 3VU University Medical Centre, Department of Gastroenterology and Hepatology, Amsterdam, Netherlands, 4OLVG, Department of Gastroenterology and Hepatology, Amsterdam, Netherlands, 5Antonius Hospital, Department of Gastroenterology and Hepatology, Nieuwegein, Netherlands
Ulcerative colitis and Crohn's colitis are associated with an increased risk of colorectal cancer (CRC). Therefore, recently updated American Gastroenterological Association (AGA) and British Society of Gastroenterology (BSG) guidelines recommend endoscopic surveillance. As surveillance intervals differ significantly between guidelines, we assessed the difference in neoplasia yield and colonoscopic workload when either the new AGA or BSG guidelines were employed.
All IBD patients enrolled in the colonoscopic surveillance program were identified in five hospitals using the patients' medical records. Patients were stratified according to the new BSG and AGA guidelines based on risk factors present at the last surveillance colonoscopy and colonoscopic workload was calculated based on the associated intervals. Cumulative incidence of colitis-associated neoplasia (CAN), defined as low-grade dysplasia (LGD) in flat mucosa or a non-adenoma like mass, high-grade dysplasia (HGD) or CRC was compared between risk groups of either guideline using Log rank testing.
A total of 1018 patients were identified that were enrolled in the surveillance program. Employing the new BSG intervals, 204 patients would be assigned to annual surveillance (20%), 393 patients to surveillance every three years (39%) and 421 patients to surveillance every five years (41%), resulting in an average of 420 colonoscopies/year. When the new AGA intervals would be applied, 64 patients (6%) would undergo annual and 954 patients (94%) biannual surveillance, resulting in an average of 541 colonoscopies/year. Thus, implementation of the new BSG guidelines could reduce the colonoscopic workload by 22% compared to the new AGA guidelines.
Yield of CAN would be 22/204 (11%) in the high risk group (12 LGD, 6 HGD, 4 CRC) and 27/393 (7%) in the intermediate risk group (7%) (21 LGD, 0 HGD, 6 CRC) and 15/421 (4%) in the low risk group (14 LGD, 0 HGD, 1 CRC) if BSG guidelines had been applied (p = 0.26). If AGA guidelines had been applied, the yield of CAN would be 13/64 (20%) in high risk group (8 LGD, 4 HGD, 1 CRC) and 51/954 (5%) in the low risk group (39 LGD, 2 HGD, 10 CRC) (p = 0.02).
Implementation of risk stratification-based intervals as recommended by the new BSG guidelines reduces the colonoscopic workload substantially compared to the AGA guidelines. However, risk stratification as recommended by the AGA seems to be more effective in discriminating between high and low risk patients than the BSG guidelines.