P182 New risk score for prediction of relapse in patients with Ulcerative Colitis, Results of a prospective cohort study
S.V. Hosseini1, P. Jafari2, S.A. Taghavi3, A. Rezaianzadeh4, M. Moini3, M. Mehrabi3, A.R. Safarpour*3
1Shiraz University of Medical Sciences, Colorectal Research Center, Shiraz, Iran, Islamic Republic of, 2Shiraz University of Medical Sciences, Biostatistics ward, Shiraz, Iran, Islamic Republic of, 3Shiraz University of Medical Sciences, Gastroenterohepatology Research Center, Shiraz, Iran, Islamic Republic of, 4Shiraz University of Medical Sciences, Nutrition and Health School, Shiraz, Iran, Islamic Republic of
The natural clinical course of Ulcerative Colitis (UC) is characterized by episodes of relapses and remissions. The aims of this study were to develop a new risk score formula for prediction of relapses in UC patients.
One hundred fifty seven patients with diagnosis of clinically UC, in remission, were followed for 12 months or shorter if they had relapsed. Hemoglobin (Hb), erythrocyte sedimentation rate (ESR), albumin (Alb), iron, platelet, white blood cells (WBC), fecal calprotectin (FC) and Seo activity index were evaluated for each patient every 3 months. Multiple logistic regression was applied to find regression coefficients of significant predictors. Receiver operating characteristics (ROC) curve analysis was performed to find the cut of points of the variables and accuracy of new risk score to prediction of relapse.
A total 157 ulcerative colitis patients were followed between Oct 2012 and Oct 2013 in periods of three month. One hundred fifty four patients were completed the study. They were 48.7% female (mean age: 42.48 ± 11.22 years, range: 20-69 years) and 51.3% male (mean age: 41.81 ± 10.82, range: 21-83 years) with no significant differences (P=0.70). The main consume medication in 64 non-relapsing (80%) and 62 relapsing (83.8%) patients was Mesalasin (P=0.65). Mean age was 42 years, mean number of previous relapses before study, mean duration of disease and mean duration of remission before study were 3.7, 99, 46 months, respectively. Seventy four patients (33 males and 41 females, 48.1%) relapsed during 12 months follow up. Final multiple logistic regression analysis demonstrated that fecal calprotectin level, number of previous relapses (NPR), age and Seo activity index can significantly predict the relapse (p=0.001, 0.024, 0.002 and <0.001, respectively). Regression coefficients of these variables (2, 2, 1 and 4, respectively) were used to develop a new risk score formula. ROC curve analysis revealed that patients with scores equal or greater than 6.5 (from total 9 scores) are the cases with high risk of relapse in the coming one year (sensitivity=73%, specificity=98%, and AUC=0.957). Finally, the following formula was obtained:
Risk Score = (2 × FC) + (2 × NPR) + (1 × age) + (4 × Seo index)
Four predictor variables which were used in the risk score formula are simple and calculate easily. These findings help to physicians for prediction of relapse risk. High score patients must be have restricted observation, treatments and follow up.