P150. Responsiveness of endoscopic indices in the evaluation of ulcerative colitis
B.G. Levesque1,2, E.V. Loftus3, R. Panaccione4, J.W. McDonald2, G. Van Assche5, G. Zou2, L. Stitt2, R. Khanna2, W.J. Sandborn1,2, G. D'Haens2,6, M.K. Vandervoort2, B.G. Feagan2,7, 1University of California, San Diego, Division of Gastroenterology, La Jolla, United States, 2Robarts Clinical Trials Inc., Robarts Research Institute, Western University, London, Ontario, Canada, 3Mayo Clinic, Department of Internal Medicine, Rochester Minnesota, United States, 4University of Calgary, Division of Gastroenterology & Hepatology, Calgary Alberta, Canada, 5University Hospital of Gasthuisberg, Department of Gastroenterology, Leuven, Belgium, 6University of Amsterdam, Academic Medical Center, Amsterdam, Netherlands, 7Western University, Department of Epidemiology and Biostatistics, Department of Medicine, London Ontario, Canada
Central endoscopic evaluation of patients is a reliable process that decreases both enrollment bias and placebo rates in clinical trials. However the responsiveness of the current endoscopic indices (EI) in ulcerative colitis (UC) has not been fully evaluated.
Data from a randomized placebo controlled trial of Asacol® (Tillotts Pharma), conducted in patients with mildly-to-moderately active UC, was used to assess the responsiveness of the Modified Mayo Clinic Endoscopic Score (MMCS), Modified Baron Score (MBS), and the Ulcerative Colitis Endoscopic Index of Severity (UCEIS) at weeks 0 and 6 or 10. Four central readers individually evaluated 121 videos obtained in clinically changed and unchanged patients. Clinical change was defined as a ≥2 point alteration in the modified UCDAI bleeding or stool frequency subscales. Responsiveness for each index was quantified by Cohen's effect size, Guyatt's responsiveness measure, and area under the receiver-operating characteristic (ROC) curves. The strength of relationship was evaluated according to conventional thresholds whereby 0.2, 0.5, and 0.8 indicate small, moderate, and large degrees of responsiveness, respectively .
Among 121 patients, 29 were clinically unchanged and 92 were changed (Table 1).
Between weeks 0 and 6 or 10, the effect sizes and Guyatt's responsiveness statistics (95% confidence intervals [CIs]) based on mean scores for the MMCS, MBS, and UCEIS were 0.49 (0.28, 0.71), 0.49 (0.28, 0.71), and 0.58 (0.36, 0.81) and 0.32 (0.11, 0.53), 0.33 (0.12, 0.54), and 0.47 (0.25, 0.69), respectively. The corresponding estimates (95% CI) for the areas under the ROC curves were 0.66 (0.55, 0.78), 0.65 (0.54, 0.77), and 0.68 (0.58, 0.79) The ROC curves are displayed in Figure 1.
|Clinically changed (n = 92)||Clinically unchanged (n = 26)||Difference (95% CI)||P value|
|MMCS||0.35||−0.35||0.70 (0.21, 1.20)||0.006|
|MBS||0.45||−0.35||0.80 (0.18, 1.43)||0.012|
|UCEIS||1.05||−0.60||1.64 (0.54, 2.74)||0.004|
Although the UCEIS had greater numerical values, the MMCS, MBS, and UCEIS displayed similar, small-to-moderate, responsiveness for the assessment of UC disease activity. These results have important implications for sample size in early drug development.
1. Husted JA, Cook RJ, Farewell VT, Gladman DD. Methods for assessing responsiveness: a critical review and recommendations. J Clin Epidemiol. 2000 May; 53(5): 459–68.