IDDF2020-ABS-0147 Development of a validated nomogram to predict aggressive Crohn's disease: a retrospective cohort study. (18th November 2020)

• Record Type: Journal Article
• Authors: Yao, Jiayin; Zhao, Junzhang; Hu, Bang; Zhi, Min
• Is Part Of: Gut. Volume 69(2020)Supplement 2
• Page Start: A49
• Page End: A50
• Publication Date: 2020-11-18
• Subjects: Gastroenterology -- Periodicals; 616.33
• DOI: 10.1136/gutjnl-2020-IDDF.93
• Languages: English
• ISSNs: 0017-5749

Abstract

Abstract : Background: Predicting aggressive Crohn's disease (CD) is crucial for determining therapeutic strategies. We aimed to develop a prognostic model to predict disease-related complications leading to early-onset surgery within 1 year after diagnosis of CD and to create a nomogram to facilitate clinical decision-making. Methods: This retrospective study was conducted from January 1, 2012, to December 31, 2016, in a single tertiary referral center, using data from patients newly diagnosed with CD and showing B1 behavior according to Montreal classification. The model was established using multivariable logistic regression analysis with evaluation of the receiver operating characteristic (ROC) curves and areas under the curve (AUC). The model was calibrated and assessed for discrimination. Further, a user-friendly nomogram was created. Results: The mean follow-up period was 53.45±12.81 months. Of 614 eligible patients, 13.5% developed surgery-related complications, including stenosis, perforation, and severe gastrointestinal bleeding. We identified age (Odds ratio (OR) 0.914, P=0.004), disease duration (OR 2.675, P<0.001), perianal disease (OR 16.013, P<0.001), previous surgery (OR 3.652, P=0.003), and extraintestinal manifestations (OR 7.625, P=0.001) as significant independent factors associated with early-onset complications and developed a prognostic model ((figure 1A ), A Prognostic model predicting complications leading to surgery within 1 year after diagnosis), whose predictive ability was appraised with AUC of 0.965, specificity of 96.71%, and sensitivity of 67.24%. This model was validated with good discrimination (AUC of 0.933), and excellent calibration was demonstrated using the Hosmer-Lemeshow goodness-of-fit test ((figure 1B ), Hosmer-Lemeshow goodness-of-fit test demonstrating a good fit of this model). A nomogram was created to facilitate clinical bedside practice ((figure 1C ) A nomogram predicting complications leading to surgery within 1 year after diagnosis in Crohn's disease patients). Conclusions: This validated prognostic model can effectively predict early-onset complications leading to surgery and screen aggressive CD, enabling physicians to customize therapeutic strategies and monitor the intensive disease.