Predicting severe inflammatory bowel disease: a risk matrix model based on the IBSEN III inception cohort.

Abstract

Background and aims: Identifying patients at risk of developing severe inflammatory bowel disease (IBD) can aid treatment decisions. However, predicting disease course remains challenging. We aimed to identify predictive factors associated with severe disease course in the first year after IBD diagnosis.

Methods: Newly diagnosed adult (≥18 years) patients with IBD were recruited from a population-based inception cohort (IBSEN III study). Preselected baseline factors were tested for associations with severe disease course, defined as IBD-related hospitalisation, surgery, treatment with ≥2 steroid courses, >2 biologics and/or new event of complication (stricture, fistula, abscess only applicable Crohn's disease (CD)). From a best fitted multivariable logistic regression model stratified by diagnosis and age (18-40/>40 years), probability of severe disease for given combinations of predictive factors was summarised in prediction matrices.

Results: At one-year follow-up, 90/559 (16%) patients with ulcerative colitis (UC) and 74/312 (24%) with CD had severe disease. Treatment with systemic steroids, vitamin D deficiency, Simple Clinical Colitis Activity Index >2, and hypoalbuminemia at diagnosis were all significantly associated with severe disease in UC patients. CD patients with stricturing or penetrating disease behaviour, systemic steroids and hypoalbuminemia at diagnosis were associated with severe disease course. The least favourable combination of these factors increased the probability of severe disease from 3% (95%CI[0-5%]) to 72% (95%CI[66-79%]) for UC and from 8% (95%CI[3-13%]) to 88% (95%CI[84-93%]) for CD.

Conclusions: Our study identified predictive factors associated with severe disease the first year after diagnosis. The probability of severe disease summarised in matrices enables easy risk stratification.