Agreement of site and central readings of ileocolonoscopic scores in Crohn's disease: comparison using data from the EXTEND trial

Abstract

Centralized endoscopic scoring may reduce variability, but evidence is lacking in patients with Crohn's disease. We assessed the agreement of endoscopic scorings between site endoscopists and one central reader by using data from the adalimumab Crohn's disease clinical trial EXTEND. Agreement between readers for Crohn's Disease Endoscopic Index of Severity (CDEIS)-scored endoscopies from 6 sites and Simple Endoscopic Score for Crohn's Disease (SES-CD)-scored endoscopies from 19 sites in EXTEND was evaluated at baseline and weeks 12 and 52. Agreement on total scores was calculated by using intraclass correlation coefficient (ICC). Kappa statistic or Spearman correlation coefficient measured the agreement between readers for each ileocolonic segment on CDEIS variables including deep ulceration, surface involved, and ulcerated surface and SES-CD variables including ulcerated surface, size of ulcers, and affected surface. ICCs on mean scores at baseline and weeks 12 and 52 were 0.78, 0.92, and 0.86 (CDEIS), and 0.77, 0.86, and 0.82 (SES-CD), respectively. Site endoscopists consistently reported higher scores. High agreement was observed for most segments and all time points for CDEIS variables and SES-CD large ulcers. Weak agreement occurred for the right side of the colon at all time points for CDEIS deep ulceration and SES-CD large ulcers and at baseline and week 12 for CDEIS ulcerated surface. Fair/moderate agreement occurred for SES-CD ulcerated surface and moderate/high agreement for affected surface for all segments and time points. Site and central readers showed high agreement on total CDEIS and SES-CD scores overall, whereas variability for individual segments was observed. Weakest agreement occurred at baseline, with a greater difference for SES-CD than for CDEIS score. ( NCT00348283.).