BACKGROUND: Identification of children with Crohn's disease (CD) at high risk of disabling disease would be invaluable in guiding initial therapy. Our study aimed to identify predictors at diagnosis of a subsequent disabling course in a population-based cohort of patients with pediatric-onset CD. METHODS: Among 537 patients with pediatric CD diagnosed at <17 years of age, 309 (57%) with 5-year follow-up were included. Clinical and demographic factors associated with subsequent disabling CD were studied. Three definitions of disabling CD were used: Saint-Antoine and Liège Hospitals' definitions and a new pediatric definition based on the presence at maximal follow-up of: 1) growth delay defined by body mass index (BMI), weight or height lower than -2 SD Z score; and 2) at least one intestinal resection or two anal interventions. Predictors were determined using multivariate analyses and their accuracy using the kappa method considering a relevant value ≥ 0.6. RESULTS: According to the Saint-Antoine definition, the rate of disabling CD was 77% and predictors were complicated behavior and L1 location. According to the Liège definition, the rate was 37% and predictors included behavior, upper gastrointestinal disease, and extraintestinal manifestations. According to the pediatric definition, the rate of disabling CD was 15%, and predictors included complicated behavior, age <14, and growth delay at diagnosis. Kappa values for each combination of predictors were, respectively, 0.2, 0.3, and 0.2 and were nonrelevant. CONCLUSIONS: Clinical parameters at diagnosis are insufficient to predict a disabling course of pediatric CD. More complex models including serological and genetic biomarkers should be tested.
BACKGROUND: Identification of children with Crohn's disease (CD) at high risk of disabling disease would be invaluable in guiding initial therapy. Our study aimed to identify predictors at diagnosis of a subsequent disabling course in a population-based cohort of patients with pediatric-onset CD. METHODS: Among 537 patients with pediatric CD diagnosed at <17 years of age, 309 (57%) with 5-year follow-up were included. Clinical and demographic factors associated with subsequent disabling CD were studied. Three definitions of disabling CD were used: Saint-Antoine and Liège Hospitals' definitions and a new pediatric definition based on the presence at maximal follow-up of: 1) growth delay defined by body mass index (BMI), weight or height lower than -2 SD Z score; and 2) at least one intestinal resection or two anal interventions. Predictors were determined using multivariate analyses and their accuracy using the kappa method considering a relevant value ≥ 0.6. RESULTS: According to the Saint-Antoine definition, the rate of disabling CD was 77% and predictors were complicated behavior and L1 location. According to the Liège definition, the rate was 37% and predictors included behavior, upper gastrointestinal disease, and extraintestinal manifestations. According to the pediatric definition, the rate of disabling CD was 15%, and predictors included complicated behavior, age <14, and growth delay at diagnosis. Kappa values for each combination of predictors were, respectively, 0.2, 0.3, and 0.2 and were nonrelevant. CONCLUSIONS: Clinical parameters at diagnosis are insufficient to predict a disabling course of pediatric CD. More complex models including serological and genetic biomarkers should be tested.
Authors: Millie D Long; Susan Hutfless; Michael D Kappelman; Hamed Khalili; Gilaad G Kaplan; Charles N Bernstein; Jean Frederic Colombel; Corinne Gower-Rousseau; Lisa Herrinton; Fernando Velayos; Edward V Loftus; Geoffrey C Nguyen; Ashwin N Ananthakrishnan; Amnon Sonnenberg; Andrew Chan; Robert S Sandler; Ashish Atreja; Samir A Shah; Kenneth J Rothman; Neal S Leleiko; Renee Bright; Paolo Boffetta; Kelly D Myers; Bruce E Sands Journal: Inflamm Bowel Dis Date: 2014-02 Impact factor: 5.325