P758 Ileorectal anastomosis vs. ileal pouch-anal anastomosis for the surgical treatment of ulcerative colitis: A Markov decision analysis
A. De Buck Van Overstraeten1,2,3, M. Brar1,2,3, S. Khorasani1,3, F. Dossa3, P. Myrelid4,5
1Mount Sinai Hospital, Department of Surgery, Toronto, Canada, 2Zane Cohen Centre for Digestive Diseases, Inflammatory Bowel Diseases, Toronto, Canada, 3University of Toronto, Department of Surgery, Toronto, Canada, 4Linkoping University Hospital, Department of Surgery, Linkoping, Sweden, 5Linkoping University, Department of Experimental Medicine, Linkoping, Sweden
Background
Ileorectal anastomosis (IRA) in patients with ulcerative colitis (UC) results in decreased postoperative morbidity and better functional outcome but leads to increased risk for rectal cancer compared with ileal pouch-anal anastomosis (IPAA). This study aims to compare IRA with IPAA in UC, using decision analysis.
Methods
A Markov simulation model was designed to simulate clinical events of IRA and IPAA over a time horizon of 40 years with time cycles of 1 year (Figure 1). The base case was a 35-year-old patient with ulcerative colitis and relatively preserved rectum. Probabilities and utilities, required to populate the model, were derived from observational studies, identified after a systematic literature search using MEDLINE. Primary outcomes were life years (LY) and quality-adjusted life years (QALY). Deterministic sensitivity analyses were performed to assess the impact of changing variables on the preferred treatment. Monte Carlo probabilistic sensitivity analysis, using 10 000 samples, was performed to account for uncertainty of variables. Prevalence of rectal cancer, IPAA and IRA failure and the stoma rate were calculated at the end of the time horizon, using markov cohort analysis.
Results
The model resulted in lower LY (36.22 vs. 37.02) and higher QALYs (33.42 vs. 31.57) for IRA. The results of the Monte Carlo probabilistic sensitivity analysis demonstrated that IRA was the preferred treatment option in 63% of the samples, accounting for a clinical significant margin of 0.25 QALY (Figure 2). The model was also sensitive to the utility of IRA, IPAA and end-ileostomy. A higher proportion of IRA patients will develop rectal cancer (7.6% vs. 3.2%) and 43.5% of all IRA patients will end with an ileostomy as opposed to 23.0% of all IPAA patients. The study was limited by characteristics inherent to modeling studies, including assumptions necessary to build the model, data input based on best available but often limited evidence and unavoidable extra- and interpolation of data.
Conclusion
IRA was the preferred treatment option when quality-adjusted life years was the outcome, with higher life years for IPAA. This model highlights that both surgical strategies are useful in ulcerative colitis patients with relatively spared rectum.