Lewis JD¹, Rutgeerts P², Feagan BG³, D'haens G⁴, Danese S⁵, Colombel JF⁶, Reinisch W⁷, Rubin DT⁸, Selinger C⁹, Bewtra M¹, Barcomb L¹⁰, Lacerda AP¹⁰, Wallace K¹⁰, Butler JW¹⁰, Wu M¹⁰, Zhou Q¹⁰, Liao X¹⁰, Sandborn WJ¹¹. Inflamm Bowel Dis. 2019 Oct 23. pii: izz241. doi: 10.1093/ibd/izz241. [Epub ahead of print]

Author information

1 Center for Clinical Epidemiology and Biostatistics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Division of Gastroenterology, University Hospital Gasthuisberg, Leuven, Belgium.

3 Department of Medicine, University of Western Ontario, London, Ontario, Canada.

4 Academic Medical Center, Amsterdam, the Netherlands.

5 Department of Gastroenterology, Istituto Clinico Humanitas, Rozzano, Milan, Italy.

6 Icahn School of Medicine at Mt Sinai, New York, New York, USA.

7 Division of Gastroenterology & Hepatology, Medical University of Vienna, Vienna, Austria.

8 University of Chicago Inflammatory Bowel Disease Center, Chicago, Illinois, USA.

9 Leeds Gastroenterology Institute, St James University Hospital, Leeds, United Kingdom.

10 AbbVie Inc., North Chicago, Illinois, USA.

11 Division of Gastroenterology, University of California San Diego, La Jolla, California, USA.

Abstract

BACKGROUND: The Crohn's Disease Activity Index (CDAI) has been criticized for being weakly correlated with bowel inflammation. We assessed correlation between Simple Endoscopic Score for Crohn's Disease (SES-CD) and individual CDAI items stratified by disease location to better understand this relationship.

METHODS: We pooled patient-level data from 3 placebo-controlled Crohn's disease (CD) trials that tested adalimumab, upadacitinib, and risankizumab. Disease location was defined as ileum only, colon only, or ileocolonic based upon colonoscopy at study entry. Pearson correlation coefficients and linear regression assessed correlations between items of the CDAI and SES-CD.

RESULTS: A total of 353 patients were included (20.7% ileal, 30.6% colonic, 48.7% ileocolonic disease). Crohn's Disease Activity Index and SES-CD scores were moderately correlated (R = 0.33; P < 0.001). Among CDAI items, the strongest correlations with SES-CD were seen with very soft or liquid stool frequency (SF) and patient-reported outcome 2 (PRO2; which includes SF and abdominal pain score; both R = 0.36; P < 0.001); these correlations were numerically stronger in colonic disease (SF: R = 0.46; P < 0.001; PRO2: R = 0.44; P < 0.001) than in ileal disease (SF: R = 0.14; P = 0.23; PRO2: R = 0.21; P = 0.07), although a test for interaction was not significant. In adjusted linear regression models, the proportion of mucosa that was inflamed and the proportion of mucosa with ulceration were positively correlated, whereas the presence of strictures was inversely correlated with SF.

CONCLUSIONS: The SF item of the CDAI is moderately correlated with SES-CD and independently correlated with mucosal ulceration, inflammation, and strictures. Understanding why bowel inflammation as measured endoscopically does not correlate more strongly with patients' symptoms could help develop scales that link CD pathology to patient experience.