Abstract
Abstract
Background
Histologic grading of severity has become an important evaluative method for ulcerative colitis (UC) early drug development clinical trials. However, its use to prognosticate patient outcome in clinical practice has not been properly validated. We therefore aimed to validate the recently developed Robarts Histological Index (RHI) of severity as a predictive outcome measure for UC through a retrospective cohort of patients.
Methods
A single gastroenterology pathologist examined and scored all baseline histopathology slides of patients with confirmed UC using the Robarts Histopathology Index (RHI) of severity. Colectomy, hospitalisation and colorectal cancer (CRC) occurrence within 5 years of diagnosis were considered major outcomes.
Results
97 patients with UC were identified, 24 patients were excluded due to incomplete data or misdiagnosis; 52 patients had the most complete follow-up data. Average age was 29.9 (+-15.7) and average duration of follow-up was 13.7 months (+-10.4). Sixty percent of patients were males and 23.5% smokers. Mean baseline RHI score was 19.5 (+-7.9) and 11.8% reported a family history of inflammatory bowel disease. 7/52 (13.5%), 27/52 (50%), and 4/52 (7.7%) underwent colectomy, required hospitalisation and developed CRC during follow-up, respectively. Stepwise model selection, identified baseline RHI score (p = 0.019), gender (p = 0.039), age (p = 0.002), and infliximab use (p = <0.001) as significant predictors of colectomy; smoking (p = 0.002), and age (p = 0.046) as significant predictors of hospitalisation; and baseline RHI score (p = <0.001), age (p = <0.001), and infliximab use (p = <0.001) as significant predictors of CRC.
Conclusions
97 patients with UC were identified, 24 patients were excluded due to incomplete data or misdiagnosis; 52 patients had the most complete follow-up data. Average age was 29.9 (+-15.7) and average duration of follow-up was 13.7 months (+-10.4). Sixty percent of patients were males and 23.5% smokers. Mean baseline RHI score was 19.5 (+-7.9) and 11.8% reported a family history of inflammatory bowel disease. 7/52 (13.5%), 27/52 (50%), and 4/52 (7.7%) underwent colectomy, required hospitalisation and developed CRC during follow-up, respectively. Stepwise model selection, identified baseline RHI score (p = 0.019), gender (p = 0.039), age (p = 0.002), and infliximab use (p = <0.001) as significant predictors of colectomy; smoking (p = 0.002), and age (p = 0.046) as significant predictors of hospitalisation; and baseline RHI score (p = <0.001), age (p = <0.001), and infliximab use (p = <0.001) as significant predictors of CRC.